Molecular Mechanisms of Glioblastoma Pathogenesis: Tusc2 Downregulation and Tgli1 Overexpression Promote Glioblastoma Development and Progression

MOLECULAR MECHANISMS OF GLIOBLASTOMA PATHOGENESIS: TUSC2 DOWNREGULATION AND TGLI1 OVEREXPRESSION PROMOTE GLIOBLASTOMA DEVELOPMENT AND PROGRESSION

TADAS KAZIMIERAS RIMKUS

A Dissertation Submitted to the Graduate Faculty of

WAKE FOREST UNIVERSITY GRADUATES SCHOOL OF ARTS AND SCIENCES

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY

Cancer Biology

May, 2019

Winston-Salem, North Carolina

Approved By:

Hui-Wen Lo, Ph.D., Advisor

Douglas S. Lyles, Ph.D., Chair

Waldemar Debinski, M.D., Ph.D.

William H. Gmeiner, Ph.D., MBA

Steven J. Kridel, Ph.D.

ACKNOWLEDGEMENTS

I would like to thank all the current and former members of the Lo Lab: Reeve Aguayo,

Ashley Anderson, Marlyn Anguelov, Austin Arrigo, Richard Carpenter, Daniel Doheny,

Alexandria Harrison, Sontoria King, Allison Lange, Adrianna Masters, Ivy Paw, Angelina

Regua, Sherona Sirkisoon, Dongqin Zhu, and most importantly the lab matriarch, Hui-

Wen Lo, for their guidance, encouragement, and friendship during my time at Wake

Forest. Good science doesn’t occur in isolation, so I want to thank all of our collaborators in the Cancer Biology department, the Wake Forest Comprehensive Cancer Center, and those at other institutions.

I would also like to thank my committee members for their invaluable advice and expertise, for challenging me to become a stronger scientist, and for being admirable scientific role models.

I want to thank all the people I’ve met here in Winston-Salem that have had an influence on me. The friendships I formed while here at Wake Forest will stay with me for all my years to come. Sarah Ewin, John Martinez, Molly McGinnis, Emily Rogers, Jamie Rose,

Eric Routh, Anirudh Sattiraju, Zach Zabarsky, and Kip Zimmerman, each of your friendships is invaluable to me and all the experiences we shared have helped me grow in countless ways, both personally and professionally. To all my friends outside of the

Wake Forest Graduate School, thank you for helping me see that life exists outside of the walls of academia.

I want to thank my dogs Cooper and Penny for being constant reminders to wake up and make the most of my day. I have never had a dull day putting up with your eccentricities, and even though you cover everything I own in your fur, I wouldn’t have it any other way.

I would also like to thank my parents, Bernadette Tutinas and Linas Rimkus, whose constant encouragement has always driven me to do my best. Your steadfast devotion to my education has brought me to where I am today and I cannot thank you enough for that.

Lastly, I would like to thank my fiancée, Deborah Luessen. You are the most amazing partner that I could ask to join me on this journey and I could not imagine my life without you. I am constantly inspired by your work ethic and scientific ability. You have made me a better scientist and person, and I can only hope to keep up with you as we grow old together. Your steadfast love for our dogs and passion for life and learning are something everyone should strive to live up to. Whether it’s our late-night life talks or binge-watching our favorite TV shows, every moment spent with you has helped shape me. Thank you for standing with me through thick and thin. I cannot wait to marry you and spend the rest of our days together.

TABLE OF CONTENTS

List of Figures and Tables………………………………………………………………………7

List of Abbreviations……………………………………………………………………………10

Abstract………………………………………………………………………………………….12

CHAPTER I

General Introduction

I.i Pathobiology and Treatment of Glioblastoma………………………………………15

I.ii Cancer Stem Cells and Glioblastoma Recurrence…………………………………16

I.iii Hedgehog Signaling Pathway………………………………………………………..17

I.iv Sonic Hedgehog Signaling in Cancer……………………………………………….20

I.v Truncated GLI1 (tGLI1) in Cancer…………………………………………………...21

I.vi Tumor Suppressor Candidate 2 (TUSC2) in Human Cancer……………………..22

I.vii Cellular Functions of TUSC2…………………………………………………………24

I.viii TUSC2 Restoration as Cancer Therapy…………………………………………….26

CHAPTER II

Truncated glioma-associated oncogene homolog 1 (tGLI1) mediates mesenchymal

glioblastoma via transcriptional activation of CD44

II.i Abstract…………………………………………………………………………………40

II.ii Introduction……………………………………………………………………………..41

II.iii Materials and Methods………………………………………………………………..43

II.iv Results………………………………………………………………………………….46

II.v Discussion………………………………………………………………………………52

II.vi Figures………………………………………………………………………………….56

II.vii Tables…………………………………………………………………………………...68

II.viii Supplementary Materials and Methods……………………………………………..69

II.ix Acknowledgements……………………………………………………………………71

II.x References……………………………………………………………………………..72

CHAPTER III

NEDD4-mediated degradation of TUSC2 promotes GBM development and progression

III.i Abstract…………………………………………………………………………………77

III.ii Introduction……………………………………………………………………………..78

III.iii Materials and Methods………………………………………………………………..81

III.iv Results………………………………………………………………………………….85

III.v Discussion………………………………………………………………………………93

III.vi Figures………………………………………………………………………………….97

III.vii Supplementary Materials and Methods……………………………………………115

III.viii Acknowledgements…………………………………………………………………..117

III.viv References……………………………………………………………………………118

CHAPTER IV

TUSC2-downregulation and tGLI1-overexpression converge to promote GBM development and progression

IV.i Materials and Methods………………………………………………………………123

IV.ii Results………………………………………………………………………………..126

IV.iii Discussion………………………………………………………………………….....129

IV.iv Figures………………………………………………………………………………...131

IV.v Acknowledgements……………………………………………………………….….136

IV.vi References……………………………………………………………………………137

CHAPTER V

General Discussion and Future Directions

V.i TUSC2 and tGLI1 as Elements in Translational Cancer Research…………….140

V.ii Future Directions……………………………………………………………………..147

V.iii Conclusion…………………………………………………………………………….151

CHAPTER VI

Appendix

VI.i TUSC2 CRISPR RNA-seq list of significantly upregulated/downregulated genes…………………………………………………………………………………………...157

CHAPTER VII

Curriculum Vitae

VII.i Curriculum Vitae………………………………………………………………...……198

LIST OF FIGURES AND TABLES

CHAPTER I

Figure 1. Canonical SHH-GLI1 signaling pathway…………………………………………19

Figure 2. TUSC2 regulation of cellular processes and signaling pathways……………..26

CHAPTER II

Figure 1. tGLI1 is expressed in a tumor-specific fashion and promotes intracranial GBM growth……………………………………………………………………………………………56

Figure 2. High tGAS is associated with poor overall survival of GBM patients and increased angiogenesis of GBM samples…………………………………………………...58

Figure 3. GBMs with high tGAS are enriched with the mesenchymal GBM gene signatures……………………………………………………………………………………….60

Figure 4. tGLI1 promotes neurosphere formation and transcriptionally activates CD44 expression……………………………………………………………………………………….62

Figure 5. tGLI1 is preferentially expressed and activated in the mesenchymal subtype of

GSC……………………………………………………………………………………………...64

Figure 6. Increased tGLI1 expression enhanced the propensity of GSC to form xenografts……………………………………………………………………………………….66

Supplemental Figure 1. CD44 Promoter Regions for tGLI1 ChIP……………………….67

Supplemental Table I. Primers for RT-qPCR………………………………………………68

Supplemental Table II. Primers for tGLI1 ChIP on CD44 Promoter……………………..68

CHAPTER III

Figure 1. TUSC2 protein is expressed in normal brain cells, but lost in GBM cells and patient samples…………………………………………………………………………………97

Figure 2. TUSC2 protein is preferentially degraded in GBM, and is bound to the E3 ligase NEDD4 in GBM cells…………………………………………………………………...99

Figure 3. TUSC2 protein is polyubiquitinated and degraded by NEDD4……………….101

Figure 4. Inverse expression of TUSC2 and NEDD4 in GBM and normal brain samples………………………………………………………………………………………...103

Figure 5. TUSC2 acts as a tumor suppressor in GBM…………………………………...105

Figure 6. TUSC2 downregulation promotes GBM growth in vitro and in vivo…………107

Figure 7. TUSC2 modulation alters expression of apoptotic proteins in GBM cells…..109

Supplemental Figure 1. Mass spectrometry analysis of TUSC2 Lysine 84 and 93 ubiquitination signals………………………………………………………………………….111

Supplemental Figure 2. TUSC2 re-expression inhibits orthotopic glioma stem cell xenograft growth………………………………………………………………………………112

Supplemental Figure 3. TCGA GBM data analysis using TUSC2 CRISPR and NEDD4 gene signatures……………………………………………………………………………….113

CHAPTER IV

Figure 1. TUSC2 downregulation alone does not induce malignant transformation….131

Figure 2. A bioinformatics approach identifies multiple oncogenic pathways enriched in

TUSC2 low GBM patients……………………………………………………………………132

Figure 3. TUSC2-downregulation and tGLI1-overexpression coordinate to transform astrocytes………………………………………………………………………………...... 133

Figure 4. TUSC2-downregulation and tGLI1-overexpression promotes GBM sphere growth…………………………………………………………………………………………..134

Supplemental Figure 1. TUSC2-downregulation and tGLI1-overexpression promotes

GBM sphere growth…………………………………………………………………………..135

LIST OF ABBREVIATIONS

ANOVA Analysis of Variance

ATCC American Type Culture Collection

Bcl-2 B-cell Lymphoma Protein 2

BTCOE Brain Tumor Center of Excellence

Ca2+ Calcium

CSC Cancer Stem Cell

DNA Deoxyribonucleic Acid

DOTAP N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium chloride

Dox Doxycycline

FACS Fluorescence-Activated Cell Soring

G1 Gap 1 phase of cell cycle

GBM Glioblastoma

GEO Gene Expression Omnibus

GLI1 Glioma-Associated Oncogene Homolog 1 gRNA Guide Ribonucleic Acid

GSC Glioma Stem Cell

GSEA Gene Set Enrichment Analysis

LNA Locked Nucleic Acid Oligonucleotide mAb Monoclonal Antibody miRNA Micro Ribonucleic Acid (aka miR)

MPM Malignant Pleural Mesothelioma mRNA Messenger Ribonucleic Acid

MTD Maximally Tolerated Dose

NEDD4 Neural Precursor Cell Expressed, Developmentally Downregulated Protein 4

NK cell Natural Killer Cell

NSCLC Non-small cell lung cancer

Olig2 Oligodendrocyte Transcription Factor 2

PBS Phosphate Buffered Saline

RFP Red Fluorescence Protein

ROS Reactive Oxygen Species

RTK Receptor Tyrosine Kinase

RT-qPCR Reverse Transcriptase Quantitative Polymerase Chain Reaction

SDS-PAGE Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis shRNA Short Hairpin Ribonucleic Acid siRNA Small Interfering Ribonucleic Acid

TCGA The Cancer Genome Atlas tGAS tGLI1 Activation Signature tGLI1 Truncated Glioma-Associated Oncogene Homolog 1

Th Helper T cells

TUNEL Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling

TUSC2 Tumor Suppressor Candidate 2

ABSTRACT

The molecular mechanisms underlying glioblastoma (GBM) development and progression have not been fully elucidated. Some of the earliest and most frequent genomic abnormalities in some cancers occur in distinct regions of chromosome 3p, suggesting these aberrations play a distinct role in the pathogenesis of these cancers.

Tumor Suppressor Candidate 2 (TUSC2, also known as FUS1) is a putative tumor suppressor first characterized in lung cancer; however, its role in GBM remains poorly understood. The goal of my thesis study was therefore to determine the role of TUSC2 in gliomagenesis and GBM progression. We found that TUSC2 protein, but not mRNA, is decreased in GBM cells and patient samples compared to normal brain tissues.

Decreased TUSC2 protein expression in GBM is due to shortened protein half-life, attributed to proteasome-dependent degradation. We discovered that the E3 ubiquitin ligase NEDD4 binds to and ubiquitinates TUSC2 at lysine 71 in GBM cells, mediating

TUSC2 protein degradation. Re-expression of TUSC2 induced GBM apoptosis in vitro and in vivo and decreased GBM incidence and growth in an orthotopic GBM xenograft mouse study. Conversely, knockout of TUSC2 in TUSC2-expressing GBM cells promoted GBM growth in vitro and in vivo. Interestingly, previous studies have shown that NEDD4 acts as a novel regulator of the Hedgehog signaling. The GLI1 transcription factor functions as the nuclear effector of the Hedgehog signaling pathway and has been widely studied as a driver of oncogenesis, tumor growth, and cancer stem cell maintenance in a variety of cancers, including GBM. Truncated GLI1 (tGLI1), is a tumor- tissue specific, gain-of-function GLI1 transcription factor isoform that directly upregulates transcription of genes promoting a highly malignant GBM phenotype both in vitro and in vivo. We showed that tGLI1 expression promotes glioma stem cell function in mesenchymal GBM cells by upregulating CD44. Given the important role that TUSC2 loss and tGLI1 gain individually plays in GBM progression, we then asked whether there

12 was a functional relationship between TUSC2 and tGLI1. Of note, whether TUSC2 and tGLI1 functionally cooperate has never been studied in any cancer or cell type. Our results showed that TUSC2 protein loss combined with high tGLI1 expression is a frequent event in GBM. TUSC2-downregulation and tGLI1-overexpression functionally cooperate to drive gliomagenesis and GBM progression partially by promoting glioma stem cell self-renewal and suppressing apoptosis. Collectively, studies conducted in my thesis projects provide novel insights into pathways that contribute to GBM development and progression.

CHAPTER I

GENERAL INTRODUCTION

Tadas K. Rimkus

Pathobiology and Treatment of Glioblastoma

Glioblastoma (GBM) is one of the deadliest primary cancers in adults, associated with a median survival of only 14 months post diagnosis (1). GBM presents as a highly diffuse and invasive brain mass on radiographic scans. Histologically, GBM has distinct regions of hypercellularity and extensive microvessel formation, and is differentiated from lower grade gliomas by the presence of necrotic core areas surrounded by regions of pseudopallisading cells (2).

Over the years, extensive genomic and molecular studies have identified numerous chromosomal and genetic alterations in malignant GBMs (3-5). Comprehensive analysis of GBM patient tumors by The Cancer Genome Atlas (TCGA) project revealed a set of core signaling pathways that are commonly dysregulated in GBM, including the p53 pathway, the RB pathway, and the receptor tyrosine kinase (RTK) pathway (6,7). Seeing as the majority of GBM cases are not identified until the patients are already symptomatic, identifying the cell of origin and the temporal genomic alterations that drive

GBM formation through a process called gliomagenesis could provide a fundamental platform for understanding disease progression and developing disease-preventative therapies. Cell-based models identified multiple genes, including H-Ras, AKT, VEGF, and FOXM1B, that drive malignant transformation of immortalized astrocytes into high- grade astrocytomas, including GBM (8-11). These models provided insights into the oncogenic role of various genes; however, they are limited in their ability to determine the GBM cell of origin and the driver mutations that accumulate in these cells.

Transgenic mouse models have been developed in order to better address the process of gliomagenesis. Numerous studies have delineated how the loss of p53, RB, PTEN, and Nf1 tumor suppressors in normal mouse astrocytes and neural stem/progenitor cells promotes the development of high-grade astrocytomas; however, additional tumor

15 suppressors and oncogenes may also be involved in gliomagenesis (12-14). Seeing as over 90% of GBM cases present as de novo tumors, with no clinical or histological evidence of pre-malignant lesions, greater understanding of the genomic aberrations that lead to GBM formation may help overcome the limited clinical efficacy of approved therapies (15).

Currently in the clinic, the standard of care for patients with GBM is surgical resection of the bulk tumor mass, followed by combination treatment of ionizing radiation and the alkylating chemotherapy, temozolomide (16). The use of tumor-treating fields, electromagnetic waves that interfere with cellular mitotic machinery, has helped extend the lives of patients receiving the treatment in addition to standard therapy (17).

However, due to the highly invasive nature of the disease as well as high rates of acquired chemoresistance, tumor recurrence is nearly universal.

Cancer Stem Cells and Glioblastoma Recurrence

The cancer stem cell (CSC) hypothesis states that a rare sub-population of cells within a tumor has stem cell-like self-renewal capabilities and can give rise to heterogeneous cancer cells that make up the tumor (18). Like in non-malignant tissues, tumors are thought to maintain a clearly defined cellular hierarchy with stem cells at the top (18).

Cancer stem cells differentiate into transient amplifying, or progenitor, cells that eventually terminally differentiate into the bulk tumor cell population. The heterogeneity seen in tumors is, therefore, likely determined by both the clonal diversity of CSCs and their differentiation capacity. In the context of GBM, these cancer stem cells are typically referred to as glioma stem cells (GSCs). It has been widely accepted that these GSCs are inherently resistant to radiation therapy and chemotherapeutic agents due to activation of DNA repair pathways and increased drug efflux from the cancer cells

(19,20). Conventional treatment for GBM typically results in a transient regression of the tumor and is almost always followed by tumor recurrence. GSCs were found to be enriched in recurrent GBM samples, thus providing evidence of GSC involvement in tumor recurrence (21,22).

GSCs, as with bulk tumor cells, are regulated by a number of intrinsic and extrinsic factors, though the distinctive functional capabilities of GSCs means they are subject to and respond to a unique set of regulatory factors. GSCs enhance transcriptional programs driven by MYC, STAT3, SOX2, GLI1, and NANOG among others (23-27). A similar study identified a core set of four transcription factors in the proneural subtype of

GBM that are capable of reprogramming differentiated tumor cells into GSCs (28).

POU3F2, SOX2, SALL2, and OLIG2 are all master transcription factors active in normal neurodevelopment, suggesting that GSCs co-opt normal stem cell programs to drive their oncogenic potential. Extrinsically, GSCs also function to redirect various brain developmental pathways to maintain their undifferentiated state. Common pathways activated in GSCs by brain-resident secreted factors include Notch, NF-κB, and Wnt signaling (29-31). In addition to these pathways, GSCs have also been shown to depend on active Hedgehog signaling (23,32,33).

Hedgehog Signaling Pathway

The Hedgehog (HH) gene was discovered in 1980 by Nusslein-Volhard and Wieschaus through genetic analysis of the fruit fly Drosophila melanogaster (34). In the early 1990s, three HH gene homologs were discovered in vertebrates; Sonic Hedgehog (SHH),

Indian Hedgehog (IHH), and Desert Hedgehog (DHH) (35-37). SHH is the most essential of these ligands in its role in normal embryonic tissue development and normal tissue maintenace, and is the most widely expressed in adult tissues (38,39). It has been

17 shown that SHH is the dominant oncogenic Hedgehog ligand, as ectopic expression of

SHH was sufficient to induce basal cell carcinoma in mice (40,41). The SHH pathway is tightly regulated in most adult tissues but aberrant activation of this pathway is found in many solid and hematological cancers (42-49). Aberrant SHH signaling in human cancers is thought to account for up to 25% of human cancer deaths, clearly indicating the need for greater understanding of SHH signaling in human cancers (50).

The canonical Hedgehog pathway contains several key components, including

Hedgehog glycoproteins SHH, IHH, and DHH (51). Upon secretion, SHH glycoproteins bind to and inactivate the 12-pass transmembrane receptor Patched1 (PTCH1), which normally inhibits the activity of the 7-pass transmembrane protein Smoothened (SMO).

In the presence of SHH ligand, PTCH1 inhibition of SMO at the primary cilium is abrogated resulting in the nuclear localization of glioma-associated (GLI) transcription factors, which are the terminal effectors of SHH signaling (Figure 1) (52,53). In the absence of Hedgehog ligands, Suppressor of Fused (SUFU) negatively regulates the pathway by directly binding to GLI transcription factors and anchoring them in the cytoplasm, thus preventing the activation of GLI target genes (54-56). Cytoplasmic sequestration of GLI transcription factors by SUFU facilitates processing and degradation of GLI proteins, therefore inhibiting Shh pathway signaling (55). SUFU has also been shown to form a repressor complex leading to interaction with DNA-bound

GLI1 and suppression of GLI1-induced gene expression (57). In vertebrates, there are three GLI transcription factors (GLI1, GLI12 and GLI3). GLI1 is the only full-length transcriptional activator whereas GLI2 and GLI3 act as either a positive or negative regulators as determined by posttranscriptional and posttranslational processing (58,59).

GLI transcription factors can activate target genes that includes targets involved in HH pathway feedback (e.g., GLI1, PTCH1), proliferation (e.g., Cyclin-D1, MYC), apoptosis

(e.g., Bcl-2), angiogenesis (e.g., ANG1/2), epithelial-to-mesenchymal transition (e.g.,

SNAIL), and stem cell self-renewal (e.g., NANOG, SOX2) (60-62).

Figure 1: Canonical SHH-GLI signaling pathway. The Hedgehog signaling pathway is held inactive in the absence of SHH ligand wherein PTCH1 inhibits SMO, resulting in GLI1 protein sequestration in the cytoplasm by SUFU. In the presence of SHH (right), PTCH1 suppression of SMO is abrogated, resulting in the nuclear translocation of GLI proteins and activation of target genes that promote oncogenic properties in tumor cells.

In addition to the canonical signaling axis, there are also non-canonical pathways that activate SHH signaling. Non-canonical SHH signaling refers to either: 1) activation of signaling via PTCH1/SMO, but independent of GLI transcription factors, or 2) activation of GLI transcription factors independent of SHH ligand or PTCH1/SMO. The latter is better studied and multiple oncogenic signaling pathways that can increase GLI activity have been identified, including K-Ras, TGF-β, PI3K/AKT, and PKC-α (63-68).

Additionally, the GLI proteins have been shown to be negatively regulated by p53, PKA, and PKC-δ (69-72).

Sonic Hedgehog Signaling in Cancer

The role of dysregulated SHH signaling in cancer was first characterized by studies of basal cell nevus syndrome. Basal cell nevus syndrome, also known as Gorlin syndrome, is an autosomal dominant disorder that presents with craniofacial and skeletal abnormalities, and a notably increased risk of basal cell carcinoma and medulloblastoma

(73,74). The major breakthrough in the understanding of SHH signaling in cancer came from the discovery that mutations in PTCH1 were the cause of Gorlin syndrome, suggesting that aberrant SHH pathway activity is responsible for the development of these cancers (75,76). These findings were reinforced by the discovery of mutations of

PTCH1, SMO, and SUFU in a large percentage of spontaneous basal cell carcinomas and medulloblastomas (77,78). The tumor suppressive role of PTCH1 has been further studied in transgenic mouse models heterozygous for a PTCH1 null mutation. These mice showed the critical features of basal cell nevus syndrome, such as development of basal cell carcinomas, medulloblastomas, and rhabdomyosarcomas (76,79). Abnormal

SHH signaling is a hallmark of many cancers. It is now understood that somatic mutations in upstream pathway elements such as SMO and PTCH1 do not account for all of the dysregulated SHH signaling observed in tumors. It has been observed in multiple tumor types that SHH pathway dysregulation can also be induced in a ligand- dependent manner through enhanced SHH autocrine or paracrine signaling. This has been reported in a wide variety of cancers including pancreatic, colorectal, metastatic prostate carcinomas, and gliomas (32,80-82). Furthermore, studies in have found that tumor cells secrete SHH ligand in a paracrine fashion which stimulates production of

20 secondary growth factors by stromal cells that drive tumor angiogenesis, tumor cell proliferation, and survival (83).

The SHH pathway has also been implicated in the regulation and maintenance of cancer stem cells (CSCs). In multiple cancer types including GBM, studies have found that SHH signaling is essential for maintenance of cancer stem cells and inhibition of the SHH pathway results in decreased stem cell propagation and renewal (23,33,43,84,85).

Multiple transcription factors involved in normal stem cell function, including NANOG,

OCT4, SOX2, and BMI1, are direct transcriptional targets of the SHH-GLI1 signaling pathway (23,86-88). Recent studies have suggested that SHH signaling in CSCs in solid tumors is involved in metastatic progression and drives epithelial-mesenchymal transition of CSCs in pancreatic and colorectal cancers, providing a link between SHH signaling in regulation of normal stem cells and its role in cancer stem cell maintenance

(80,82,89).

Truncated GLI1 (tGLI1) in Cancer

Recently, a novel alternative splice variant of GLI1 was discovered in our laboratory, termed truncated GLI1 (tGLI1) (90). tGLI1 is an alternatively spliced variant of GLI1 with an in-frame deletion of 123 base pairs (i.e. 41 amino acids) which removes the entirety of exon 3 and part of exon 4, without any effect on the functional domains at the mRNA level. Evidence to date indicates that tGLI1 is expressed at high levels in GBM and breast cancer but undetectable or expressed at a low level in normal tissues. We have observed that tGLI1 promotes cell migration, invasion, and angiogenesis by upregulating

CD24, HPA1, VEGF-C, and TEM-7 in GBM and breast cancer (91-93). We have also shown that tGLI1 upregulates VEGF-A and VEGFR-2, leading to a distinct autocrine loop that promotes angiogenesis and cell growth in breast cancer (94). Overexpression

21 of tGLI1 results in larger tumor growth and greater tumor angiogenesis compared to tumors overexpressing GLI1 (90-94). These studies indicate that tGLI1 is a gain-of- function transcription factor compared to GLI1 as it has shown the ability to upregulate several genes that are not GLI1 target genes. Recently, we discovered that tGLI1 expression promotes a more stem-like phenotype in GBM cells by directly upregulating transcription of CD44 (95). Patients with mesenchymal subtype GBM, a subtype of GBM characterized by high chemo- and radio-resistance and poorer overall survival, showed higher expression of a tGLI1 activity gene signature, underlining the role that tGLI1 plays in GSCs and GBM recurrence.

Tumor Suppressor Candidate 2 (TUSC2) in Human Cancer

Tumor suppressor candidate 2 (TUSC2, also known as FUS1) was identified as a candidate tumor suppressor gene located in a region on chromosome 3p21.3 that is homozygously deleted in some lung and breast cancers (96). However, 3p21.3 deletion was later found to be very rare in lung cancer (1.1%; The Cancer Genome Atlas or

TCGA) and most cancer types, except for malignant pleural mesothelioma (36%) (97). In lung cancer, there is some evidence of increased methylation in the TUSC2 gene promoter region compared to normal tissues (98,99). Additionally, the TUSC2 gene promoter region was reported to be partially methylated in head and neck squamous cell tumors, but unmethylated in normal mucosa (100). In lung cancer, for which TUSC2 is considered as a tumor suppressor gene, loss of TUSC2 expression has been observed in between 80-100% of lung tumors and the loss is attributed to 3p21.3 deletion, loss of heterozygosity, and post-transcriptional repression by microRNAs (96,101-103). TUSC2 somatic mutations have not been found in any cancer specimens according to TCGA, albeit infrequent mutations have been reported in lung cancer cell lines (98). The TUSC2 protein consists of 110 amino acids and is located predominately in mitochondria and

22 cytoplasm (97,98). Myristoylation is the only post-translational modification of the TUSC2 protein that has been reported. Using surface-enhanced laser desorption/ionization mass spectrometric analysis on an anti-TUSC2 antibody-capture

ProteinChip array, Uno et al. identified TUSC2 as an N-myristoylated protein.

Myristoylation is required for its tumor-suppressive activity in human lung cancer cells. A myristoylation-deficient TUSC2 mutant lost the ability to inhibit cell growth and induce apoptosis in lung cancer, in vitro and in vivo, and is susceptible to rapid proteasome- dependent degradation (104).

Loss of TUSC2 mRNA expression has been observed in ~80% of lung tumors (96). Loss and reduction of TUSC2 protein expression was detected in 82% of non-small cell lung cancers (NSCLCs) and 100% of small cell lung carcinomas (103). This loss of TUSC2 expression has been shown to occur early in lung cancer pathogenesis. Bronchial squamous metaplastic and dysplastic lesions expressed lower levels of TUSC2 protein compared to normal and hyperplastic epithelia. In NSCLCs, loss of TUSC2 protein, determined by immunohistochemical staining, was associated with significantly worse overall survival (103). In malignant pleural mesothelioma (MPM), an aggressive inflammatory cancer associated with exposure to asbestos, TUSC2 mRNA was downregulated in approximately 84% of 30 MPM specimens while loss of 3p21.3 region was observed in approximately 36% of MPMs including stage 1 tumors (97). In gliomas,

TUSC2 protein expression was reported to be reduced in high-grade tumors compared to low-grade tumors; however, normal brain tissues were not examined in this study

(105). Interestingly, TUSC2 mRNA expression was detected in almost all sarcoma cell lines, benign bone and soft-tissue sarcoma tissues, and healthy tissues. In contrast,

TUSC2 protein expression was undetected in the majority of sarcoma cells and sarcoma tissues (106).

Cellular Functions of TUSC2

The exact functions of TUSC2 are still not clear. TUSC2 is associated with several pathways and cellular processes (Figure 2) (107). Forced TUSC2 expression can decrease cell growth, induce G1 cell cycle arrest, and promote apoptosis in lung cancer in vitro and in vivo (98,104,108). TUSC2-knockout mice were not embryonically lethal, but displayed complex immunological phenotype and developed spontaneous vascular and hematopoietic tumors (109). Restoration of TUSC2 expression has been explored as an anti-cancer therapy using cell lines and xenograft mouse models and in a phase I clinical trial of lung cancer (104,110-112). TUSC2 appeared to affect transcription profile of MPM (97). The authors conducted gene expression arrays with TUSC2-transfected

MPM cells and found that TUSC2 transfection led to increased expression of multiple genes with tumor suppressor properties and reduced expression of pro-tumorigenic genes.

TUSC2 has been shown to regulate protein tyrosine and serine/threonine kinases. A

TUSC2-derived peptide inhibits the activity of Abl tyrosine kinase in NSCLC cells (113).

Forced expression of full-length TUSC2 decreased levels of activated c-Abl and inhibited its tyrosine kinase activity, suggesting c-Abl is a possible target of TUSC2 in NSCLC.

Additional protein tyrosine kinases and Ser/Thr kinases inhibited by TUSC2 include

EGFR, PDGFR, AKT, c-Abl, FGFR, and c-Kit (114-117). TUSC2 can also inhibit mTOR activation (114). TUSC2 transient expression in LKB1-defective NSCLC cells significantly stimulated AMP-activated protein kinase (AMPK) phosphorylation and its enzymatic activity (117).

TUSC2 has also been shown to regulate mitochondrial calcium handling. Since TUSC2 possesses putative calcium-binding and myristoyl-binding domains, Uzhachenko et al.

24 found that TUSC2 regulates mitochondrial calcium handling and calcium-coupled processes (118). TUSC2 loss led to reduced mitochondrial calcium uptake in calcium- loaded epithelial cells, splenocytes, and activated CD4+ T cells. Ex vivo analysis of activated CD4+ T cells showed TUSC2-dependent changes in calcium-regulated processes, such as surface expression of CD4 and PD-1/PD-L1, proliferation, and Th polarization. In contrast, TUSC2-knockout T cells showed an increase of basal activation of calcium-dependent NF-κB and NFAT targets, but the loss of TUSC2 impaired T cell response to activation by CD3/CD28. (118). Since mitochondrial calcium homeostasis is an important contributing factor to premature aging and aging-associated pathologies, it was observed that TUSC2-KO mice developed multiple early aging signs including lordokyphosis, lack of vigor, inability to accumulate fat, reduced ability to tolerate stress, and premature death; as well as low sperm counts, compromised ability of adult stem cells to repopulate tissues, and chronic inflammation. In addition to altered mitochondrial calcium homeostasis, TUSC2-KO cells have low reserve respiratory capacity, suggesting that energy homeostasis is also controlled in part by TUSC2 (119-122).

Figure 2: TUSC2 regulation of cellular processes and signaling pathways. TUSC2 has been shown to regulate diverse cellular processes including cell cycle arrest, apoptosis, calcium homeostasis, oxidative stress response, immune response, miRNA expression, transcriptional regulation, p53/MDM2 pathway, and tyrosine and serine/threonine kinases.

TUSC2 Restoration as Cancer Therapy

TUSC2 restoration is an attractive strategy to inhibit tumor growth and progression given its tumor suppressive effects. Overexpression of TUSC2 has been shown to have antitumor effects in lung cancer, GBM, esophageal carcinoma, and thyroid cancer

(105,108,110-112,114,117,123-127). In NSCLC models, exogenous restoration of

TUSC2 expression sensitized wild-type EGFR-expressing lung cancer cells to EGFR inhibitors (114). The authors reported that combination treatment with intravenous

TUSC2 nanovesicles and erlotinib (an EGFR inhibitor) synergistically inhibited tumor

26 growth and metastasis, and increased apoptotic activity of NSCLC xenografts. TUSC2 inhibits mTOR activation, and TUSC2 restoration rendered tumors more responsive to the mTOR inhibitor rapamycin in combination with erlotinib. A more recent study further reported that increasing reactive oxygen species (ROS) via addition of the thioredoxin reductase 1 inhibitor auranofin enhanced therapeutic efficacy of the combination of

TUSC2 with erlotinib in wild-type EGFR-expressing lung cancer (128).

TUSC2 restoration also sensitized LKB1-deficient NSCLC to the AKT inhibitor MK2206

(117). In contrast, TUSC2 did not affect the response to MK2206 treatment in two LKB1- proficient NSCLC cell lines. Systemic TUSC2 delivery by nanoparticle gene transfer, combined with MK2206, inhibited growth of LKB1-defective lung cancer xenografts.

Interestingly, TUSC2 transient expression in LKB1-defective cells significantly activated

AMPK phosphorylation and enzymatic activity. AMPK gene knockdown abrogated the

TUSC2-MK2206 cooperation, suggesting that TUSC2 restoration is a viable strategy to overcome tumor resistance to AKT-targeted therapy (117). More recently, the same group explored the use of combined gene therapy with LKB1 and TUSC2, two tumor suppressors, as a cancer therapy for lung cancer (108). The results indicated that intratumoral administration of TUSC2-LKB1 liposomes led to inhibition of subcutaneous lung tumor xenograft while intravenous injections resulted in a reduction of metastatic tumor modules.

Co-expression of TUSC2 with another tumor suppressor p53 also synergistically suppressed lung tumor growth (123). Co-expression of TUSC2 with human IL-12 inhibits lung tumor growth and metastasis (126). This co-expression was found to induce a strong anti-tumor immune response by secreting much higher levels of interferon- gamma and IL-15, enhancing expression of MHC-I and Fas, and increasing infiltration of activated CD4+ and CD8+ T lymphocytes. TUSC2-hIL-12 co-expression could also

27 induce tumor cell apoptosis and inhibit tumor cell proliferation partly by higher activation of STAT1 signal pathway and upregulation of p53 (126). TUSC2 restoration can also potentiate the effects of cisplatin in NSCLC (124). Systemic administration of TUSC2 nanoparticle sensitized NSCLC tumors to cisplatin. TUSC2-enhanced chemosensitivity is associated with the downregulation of MDM2, accumulation of p53, and activation of the Apaf-1-dependent apoptosis pathway. Interestingly, TUSC2 is associated with radioprotection of normal tissues (129,130).

A phase I clinical study with systemically administered TUSC2 nanoparticles was completed in NSCLC patients in 2012 (NCT00059605) (111). Thirty-one patients with recurrent and/or metastatic lung cancer previously treated with platinum-based chemotherapy were enrolled in this study. Patients were treated with six escalating doses of intravenous N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium chloride

(DOTAP):cholesterol nanoparticles encapsulating a TUSC2 expression plasmid

(DOTAP:chol-TUSC2) every 3 weeks. The maximally tolerated dose, MTD, was determined to be 0.06 mg/kg. Five patients achieved stable disease (2.6-10.8 months, including 2 minor responses). The authors concluded that DOTAP:chol-TUSC2 can be safely administered intravenously in lung cancer patients. This encouraging outcome has led to ongoing phase I/II trials with stage IV lung cancer patients (NCT01455389). In the phase I trial, TUSC2-nanoparticles combined with erlotinib (EGFR inhibitor) will be tested to determine MTD. The goal of the phase II trial is to examine whether the combination of TUSC2-nanoparticles and erlotinib can help to control NSCLC.

REFERENCES

1. Johnson DR, O'Neill BP. Glioblastoma survival in the United States before and during the temozolomide era. Journal of neuro-oncology 2012;107:359-64 2. Urbańska K, Sokołowska J, Szmidt M, Sysa P. Glioblastoma multiforme - an overview. Contemporary oncology (Poznan, Poland) 2014;18:307-12 3. Gravendeel LA, Kouwenhoven MC, Gevaert O, de Rooi JJ, Stubbs AP, Duijm JE, et al. Intrinsic gene expression profiles of gliomas are a better predictor of survival than histology. Cancer Res 2009;69:9065-72 4. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 2006;9:157-73 5. Verhaak RGW, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, et al. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer cell 2010;17:98-110 6. The Cancer Genome Atlas Research N, McLendon R, Friedman A, Bigner D, Van Meir EG, Brat DJ, et al. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 2008;455:1061 7. Brennan CW, Verhaak RGW, McKenna A, Campos B, Noushmehr H, Salama SR, et al. The somatic genomic landscape of glioblastoma. Cell 2013;155:462-77 8. Dai B, Pieper RO, Li D, Wei P, Liu M, Woo SY, et al. FoxM1B regulates NEDD4- 1 expression, leading to cellular transformation and full malignant phenotype in immortalized human astrocytes. Cancer research 2010;70:2951-61 9. Sonoda Y, Kanamori M, Deen DF, Cheng SY, Berger MS, Pieper RO. Overexpression of vascular endothelial growth factor isoforms drives oxygenation and growth but not progression to glioblastoma multiforme in a human model of gliomagenesis. Cancer Res 2003;63:1962-8 10. Sonoda Y, Ozawa T, Aldape KD, Deen DF, Berger MS, Pieper RO. Akt Pathway Activation Converts Anaplastic Astrocytoma to Glioblastoma Multiforme in a Human Astrocyte Model of Glioma. Cancer Research 2001;61:6674-8 11. Sonoda Y, Ozawa T, Hirose Y, Aldape KD, McMahon M, Berger MS, et al. Formation of Intracranial Tumors by Genetically Modified Human Astrocytes Defines Four Pathways Critical in the Development of Human Anaplastic Astrocytoma. Cancer Research 2001;61:4956-60 12. Alcantara Llaguno S, Chen J, Kwon CH, Jackson EL, Li Y, Burns DK, et al. Malignant astrocytomas originate from neural stem/progenitor cells in a somatic tumor suppressor mouse model. Cancer Cell 2009;15:45-56 13. Chow Lionel ML, Endersby R, Zhu X, Rankin S, Qu C, Zhang J, et al. Cooperativity within and among Pten, p53, and Rb Pathways Induces High- Grade Astrocytoma in Adult Brain. Cancer Cell 2011;19:305-16 14. Kwon C-H, Zhao D, Chen J, Alcantara S, Li Y, Burns DK, et al. Pten Haploinsufficiency Accelerates Formation of High-Grade Astrocytomas. Cancer Research 2008;68:3286-94 15. Ohgaki H, Kleihues P. The definition of primary and secondary glioblastoma. Clinical cancer research : an official journal of the American Association for Cancer Research 2013;19:764-72 16. Davis ME. Glioblastoma: Overview of Disease and Treatment. Clinical journal of oncology nursing 2016;20:S2-S8

17. Rick J, Chandra A, Aghi MK. Tumor treating fields: a new approach to glioblastoma therapy. Journal of neuro-oncology 2018;137:447-53 18. Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer, and cancer stem cells. Nature 2001;414:105 19. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, et al. Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 2006;444:756-60 20. Bleau AM, Hambardzumyan D, Ozawa T, Fomchenko EI, Huse JT, Brennan CW, et al. PTEN/PI3K/Akt pathway regulates the side population phenotype and ABCG2 activity in glioma tumor stem-like cells. Cell stem cell 2009;4:226-35 21. Auffinger B, Tobias AL, Han Y, Lee G, Guo D, Dey M, et al. Conversion of differentiated cancer cells into cancer stem-like cells in a glioblastoma model after primary chemotherapy. Cell death and differentiation 2014;21:1119-31 22. Chen J, Li Y, Yu TS, McKay RM, Burns DK, Kernie SG, et al. A restricted cell population propagates glioblastoma growth after chemotherapy. Nature 2012;488:522-6 23. Clement V, Sanchez P, de Tribolet N, Radovanovic I, Ruiz i Altaba A. HEDGEHOG-GLI1 signaling regulates human glioma growth, cancer stem cell self-renewal, and tumorigenicity. Current biology : CB 2007;17:165-72 24. Gangemi RM, Griffero F, Marubbi D, Perera M, Capra MC, Malatesta P, et al. SOX2 silencing in glioblastoma tumor-initiating cells causes stop of proliferation and loss of tumorigenicity. Stem cells (Dayton, Ohio) 2009;27:40-8 25. Sherry MM, Reeves A, Wu JK, Cochran BH. STAT3 is required for proliferation and maintenance of multipotency in glioblastoma stem cells. Stem cells (Dayton, Ohio) 2009;27:2383-92 26. Wang J, Wang H, Li Z, Wu Q, Lathia JD, McLendon RE, et al. c-Myc is required for maintenance of glioma cancer stem cells. PloS one 2008;3:e3769-e 27. Zbinden M, Duquet A, Lorente-Trigos A, Ngwabyt SN, Borges I, Ruiz i Altaba A. NANOG regulates glioma stem cells and is essential in vivo acting in a cross- functional network with GLI1 and p53. The EMBO journal 2010;29:2659-74 28. Suvà ML, Rheinbay E, Gillespie SM, Patel AP, Wakimoto H, Rabkin SD, et al. Reconstructing and reprogramming the tumor-propagating potential of glioblastoma stem-like cells. Cell 2014;157:580-94 29. Hovinga KE, Shimizu F, Wang R, Panagiotakos G, Van Der Heijden M, Moayedpardazi H, et al. Inhibition of notch signaling in glioblastoma targets cancer stem cells via an endothelial cell intermediate. Stem cells (Dayton, Ohio) 2010;28:1019-29 30. Rheinbay E, Suvà ML, Gillespie SM, Wakimoto H, Patel AP, Shahid M, et al. An aberrant transcription factor network essential for Wnt signaling and stem cell maintenance in glioblastoma. Cell reports 2013;3:1567-79 31. Rinkenbaugh AL, Cogswell PC, Calamini B, Dunn DE, Persson AI, Weiss WA, et al. IKK/NF-κB signaling contributes to glioblastoma stem cell maintenance. Oncotarget 2016;7:69173-87 32. Bar EE, Chaudhry A, Lin A, Fan X, Schreck K, Matsui W, et al. Cyclopamine- mediated hedgehog pathway inhibition depletes stem-like cancer cells in glioblastoma. Stem cells (Dayton, Ohio) 2007;25:2524-33 33. Ehtesham M, Sarangi A, Valadez JG, Chanthaphaychith S, Becher MW, Abel TW, et al. Ligand-dependent activation of the hedgehog pathway in glioma progenitor cells. Oncogene 2007;26:5752-61 34. Nüsslein-Volhard C, Wieschaus E. Mutations affecting segment number and polarity in Drosophila. Nature 1980;287:795-801

35. Echelard Y, Epstein DJ, St-Jacques B, Shen L, Mohler J, McMahon JA, et al. Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 1993;75:1417-30 36. Krauss S, Concordet JP, Ingham PW. A functionally conserved homolog of the Drosophila segment polarity gene hh is expressed in tissues with polarizing activity in zebrafish embryos. Cell 1993;75:1431-44 37. Roelink H, Augsburger A, Heemskerk J, Korzh V, Norlin S, Ruiz i Altaba A, et al. Floor plate and motor neuron induction by vhh-1, a vertebrate homolog of hedgehog expressed by the notochord. Cell 1994;76:761-75 38. Ingham PW, McMahon AP. Hedgehog signaling in animal development: paradigms and principles. Genes & development 2001;15:3059-87 39. Pathi S, Pagan-Westphal S, Baker DP, Garber EA, Rayhorn P, Bumcrot D, et al. Comparative biological responses to human Sonic, Indian, and Desert hedgehog. Mechanisms of development 2001;106:107-17 40. Fan H, Oro AE, Scott MP, Khavari PA. Induction of basal cell carcinoma features in transgenic human skin expressing Sonic Hedgehog. Nature Medicine 1997;3:788-92 41. Oro AE, Higgins KM, Hu Z, Bonifas JM, Epstein EH, Jr., Scott MP. Basal cell carcinomas in mice overexpressing sonic hedgehog. Science (New York, NY) 1997;276:817-21 42. Berman DM, Karhadkar SS, Hallahan AR, Pritchard JI, Eberhart CG, Watkins DN, et al. Medulloblastoma growth inhibition by hedgehog pathway blockade. Science (New York, NY) 2002;297:1559-61 43. Dierks C, Beigi R, Guo GR, Zirlik K, Stegert MR, Manley P, et al. Expansion of Bcr-Abl-positive leukemic stem cells is dependent on Hedgehog pathway activation. Cancer Cell 2008;14:238-49 44. Kubo M, Nakamura M, Tasaki A, Yamanaka N, Nakashima H, Nomura M, et al. Hedgehog signaling pathway is a new therapeutic target for patients with breast cancer. Cancer Res 2004;64:6071-4 45. Sanchez P, Hernández AM, Stecca B, Kahler AJ, DeGueme AM, Barrett A, et al. Inhibition of prostate cancer proliferation by interference with SONIC HEDGEHOG-GLI1 signaling. Proceedings of the National Academy of Sciences of the United States of America 2004;101:12561-6 46. Thayer SP, di Magliano MP, Heiser PW, Nielsen CM, Roberts DJ, Lauwers GY, et al. Hedgehog is an early and late mediator of pancreatic cancer tumorigenesis. Nature 2003;425:851-6 47. Tojo M, Kiyosawa H, Iwatsuki K, Kaneko F. Expression of a sonic hedgehog signal transducer, hedgehog-interacting protein, by human basal cell carcinoma. British Journal of Dermatology 2002;146:69-73 48. Watkins DN, Berman DM, Burkholder SG, Wang B, Beachy PA, Baylin SB. Hedgehog signalling within airway epithelial progenitors and in small-cell lung cancer. Nature 2003;422:313-7 49. Yuan Z, Goetz JA, Singh S, Ogden SK, Petty WJ, Black CC, et al. Frequent requirement of hedgehog signaling in non-small cell lung carcinoma. Oncogene 2007;26:1046-55 50. Lum L, Beachy PA. The Hedgehog response network: sensors, switches, and routers. Science (New York, NY) 2004;304:1755-9 51. Varjosalo M, Taipale J. Hedgehog: functions and mechanisms. Genes & development 2008;22:2454-72 52. Carpenter D, Stone DM, Brush J, Ryan A, Armanini M, Frantz G, et al. Characterization of two patched receptors for the vertebrate hedgehog protein

family. Proceedings of the National Academy of Sciences of the United States of America 1998;95:13630-4 53. Rimkus TK, Carpenter RL, Qasem S, Chan M, Lo HW. Targeting the Sonic Hedgehog Signaling Pathway: Review of Smoothened and GLI Inhibitors. Cancers 2016;8 54. Cheng SY, Bishop JM. Suppressor of Fused represses Gli-mediated transcription by recruiting the SAP18-mSin3 corepressor complex. Proceedings of the National Academy of Sciences 2002;99:5442-7 55. Kogerman P, Grimm T, Kogerman L, Krause D, Unden AB, Sandstedt B, et al. Mammalian suppressor-of-fused modulates nuclear-cytoplasmic shuttling of Gli- 1. Nature cell biology 1999;1:312-9 56. Paces-Fessy M, Boucher D, Petit E, Paute-Briand S, Blanchet-Tournier MF. The negative regulator of Gli, Suppressor of fused (Sufu), interacts with SAP18, Galectin3 and other nuclear proteins. The Biochemical journal 2004;378:353-62 57. Chen JK, Taipale J, Cooper MK, Beachy PA. Inhibition of Hedgehog signaling by direct binding of cyclopamine to Smoothened. Genes & development 2002;16:2743-8 58. Ruiz i Altaba A. Catching a Gli-mpse of Hedgehog. Cell 1997;90:193-6 59. Sasaki H, Nishizaki Y, Hui C, Nakafuku M, Kondoh H. Regulation of Gli2 and Gli3 activities by an amino-terminal repression domain: implication of Gli2 and Gli3 as primary mediators of Shh signaling. Development (Cambridge, England) 1999;126:3915-24 60. Hui CC, Angers S. Gli proteins in development and disease. Annual review of cell and developmental biology 2011;27:513-37 61. Scales SJ, de Sauvage FJ. Mechanisms of Hedgehog pathway activation in cancer and implications for therapy. Trends in pharmacological sciences 2009;30:303-12 62. Stecca B, Ruiz I Altaba A. Context-dependent regulation of the GLI code in cancer by HEDGEHOG and non-HEDGEHOG signals. Journal of molecular cell biology 2010;2:84-95 63. Deng W, Vanderbilt DB, Lin C-C, Martin KH, Brundage KM, Ruppert JM. SOX9 inhibits β-TrCP-mediated protein degradation to promote nuclear GLI1 expression and cancer stem cell properties. Journal of cell science 2015;128:1123-38 64. Ke Z, Caiping S, Qing Z, Xiaojing W. Sonic hedgehog-Gli1 signals promote epithelial-mesenchymal transition in ovarian cancer by mediating PI3K/AKT pathway. Medical oncology (Northwood, London, England) 2015;32:368 65. Rajurkar M, De Jesus-Monge WE, Driscoll DR, Appleman VA, Huang H, Cotton JL, et al. The activity of Gli transcription factors is essential for Kras-induced pancreatic tumorigenesis. Proc Natl Acad Sci U S A 2012;109:E1038-47 66. Ramaswamy B, Lu Y, Teng KY, Nuovo G, Li X, Shapiro CL, et al. Hedgehog signaling is a novel therapeutic target in tamoxifen-resistant breast cancer aberrantly activated by PI3K/AKT pathway. Cancer Res 2012;72:5048-59 67. Seto M, Ohta M, Asaoka Y, Ikenoue T, Tada M, Miyabayashi K, et al. Regulation of the hedgehog signaling by the mitogen-activated protein kinase cascade in gastric cancer. Molecular carcinogenesis 2009;48:703-12 68. Zhou J, Zhu G, Huang J, Li L, Du Y, Gao Y, et al. Non-canonical GLI1/2 activation by PI3K/AKT signaling in renal cell carcinoma: A novel potential therapeutic target. Cancer letters 2016;370:313-23

69. Makinodan E, Marneros AG. Protein kinase A activation inhibits oncogenic Sonic hedgehog signalling and suppresses basal cell carcinoma of the skin. Experimental dermatology 2012;21:847-52 70. Sheng T, Chi S, Zhang X, Xie J. Regulation of Gli1 Localization by the cAMP/Protein Kinase A Signaling Axis through a Site Near the Nuclear Localization Signal. Journal of Biological Chemistry 2006;281:9-12 71. Stecca B, Ruiz i Altaba A. A GLI1-p53 inhibitory loop controls neural stem cell and tumour cell numbers. The EMBO journal 2009;28:663-76 72. Yoon JW, Lamm M, Iannaccone S, Higashiyama N, Leong KF, Iannaccone P, et al. p53 modulates the activity of the GLI1 oncogene through interactions with the shared coactivator TAF9. DNA repair 2015;34:9-17 73. Gorlin RJ, Goltz RW. Multiple nevoid basal-cell epithelioma, jaw cysts and bifid rib. A syndrome. The New England journal of medicine 1960;262:908-12 74. Kimonis VE, Goldstein AM, Pastakia B, Yang ML, Kase R, DiGiovanna JJ, et al. Clinical manifestations in 105 persons with nevoid basal cell carcinoma syndrome. American journal of medical genetics 1997;69:299-308 75. Aszterbaum M, Rothman A, Johnson RL, Fisher M, Xie J, Bonifas JM, et al. Identification of mutations in the human PATCHED gene in sporadic basal cell carcinomas and in patients with the basal cell nevus syndrome. The Journal of investigative dermatology 1998;110:885-8 76. Hahn H, Wicking C, Zaphiropoulous PG, Gailani MR, Shanley S, Chidambaram A, et al. Mutations of the human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Cell 1996;85:841-51 77. Epstein EH. Basal cell carcinomas: attack of the hedgehog. Nature reviews Cancer 2008;8:743-54 78. Kool M, Koster J, Bunt J, Hasselt NE, Lakeman A, van Sluis P, et al. Integrated genomics identifies five medulloblastoma subtypes with distinct genetic profiles, pathway signatures and clinicopathological features. PLoS One 2008;3:e3088 79. Goodrich LV, Milenkovic L, Higgins KM, Scott MP. Altered neural cell fates and medulloblastoma in mouse patched mutants. Science (New York, NY) 1997;277:1109-13 80. Feldmann G, Dhara S, Fendrich V, Bedja D, Beaty R, Mullendore M, et al. Blockade of hedgehog signaling inhibits pancreatic cancer invasion and metastases: a new paradigm for combination therapy in solid cancers. Cancer research 2007;67:2187-96 81. Karhadkar SS, Bova GS, Abdallah N, Dhara S, Gardner D, Maitra A, et al. Hedgehog signalling in prostate regeneration, neoplasia and metastasis. Nature 2004;431:707-12 82. Varnat F, Duquet A, Malerba M, Zbinden M, Mas C, Gervaz P, et al. Human colon cancer epithelial cells harbour active HEDGEHOG-GLI signalling that is essential for tumour growth, recurrence, metastasis and stem cell survival and expansion. EMBO molecular medicine 2009;1:338-51 83. Yauch RL, Gould SE, Scales SJ, Tang T, Tian H, Ahn CP, et al. A paracrine requirement for hedgehog signalling in cancer. Nature 2008;455:406-10 84. Liu S, Dontu G, Mantle ID, Patel S, Ahn NS, Jackson KW, et al. Hedgehog signaling and Bmi-1 regulate self-renewal of normal and malignant human mammary stem cells. Cancer Res 2006;66:6063-71 85. Zhao C, Chen A, Jamieson CH, Fereshteh M, Abrahamsson A, Blum J, et al. Hedgehog signalling is essential for maintenance of cancer stem cells in myeloid leukaemia. Nature 2009;458:776-9

86. Batsaikhan BE, Yoshikawa K, Kurita N, Iwata T, Takasu C, Kashihara H, et al. Cyclopamine decreased the expression of Sonic Hedgehog and its downstream genes in colon cancer stem cells. Anticancer research 2014;34:6339-44 87. Justilien V, Walsh MP, Ali SA, Thompson EA, Murray NR, Fields AP. The PRKCI and SOX2 oncogenes are coamplified and cooperate to activate Hedgehog signaling in lung squamous cell carcinoma. Cancer Cell 2014;25:139-51 88. Po A, Ferretti E, Miele E, De Smaele E, Paganelli A, Canettieri G, et al. Hedgehog controls neural stem cells through p53-independent regulation of Nanog. The EMBO journal 2010;29:2646-58 89. Rasheed ZA, Yang J, Wang Q, Kowalski J, Freed I, Murter C, et al. Prognostic significance of tumorigenic cells with mesenchymal features in pancreatic adenocarcinoma. Journal of the National Cancer Institute 2010;102:340-51 90. Lo H-W, Zhu H, Cao X, Aldrich A, Ali-Osman F. A novel splice variant of GLI1 that promotes glioblastoma cell migration and invasion. Cancer research 2009;69:6790-8 91. Cao X, Geradts J, Dewhirst MW, Lo HW. Upregulation of VEGF-A and CD24 gene expression by the tGLI1 transcription factor contributes to the aggressive behavior of breast cancer cells. Oncogene 2012;31:104-15 92. Carpenter RL, Paw I, Zhu H, Sirkisoon S, Xing F, Watabe K, et al. The gain-of- function GLI1 transcription factor TGLI1 enhances expression of VEGF-C and TEM7 to promote glioblastoma angiogenesis. Oncotarget 2015;6:22653-65 93. Zhu H, Carpenter RL, Han W, Lo H-W. The GLI1 splice variant TGLI1 promotes glioblastoma angiogenesis and growth. Cancer letters 2014;343:51-61 94. Han W, Carpenter RL, Lo H-W. TGLI1 Upregulates Expression of VEGFR2 and VEGF-A, Leading to a Robust VEGF-VEGFR2 Autocrine Loop and Cancer Cell Growth. Cancer Hallmarks 2013;1:28-37 95. Rimkus TK, Carpenter RL, Sirkisoon S, Zhu D, Pasche BC, Chan MD, et al. Truncated Glioma-Associated Oncogene Homolog 1 (tGLI1) Mediates Mesenchymal Glioblastoma via Transcriptional Activation of CD44. Cancer Res 2018;78:2589-600 96. Lerman MI, Minna JD. The 630-kb lung cancer homozygous deletion region on human chromosome 3p21.3: identification and evaluation of the resident candidate tumor suppressor genes. The International Lung Cancer Chromosome 3p21.3 Tumor Suppressor Gene Consortium. Cancer Res 2000;60:6116-33 97. Ivanova AV, Ivanov SV, Prudkin L, Nonaka D, Liu Z, Tsao A, et al. Mechanisms of FUS1/TUSC2 deficiency in mesothelioma and its tumorigenic transcriptional effects. Molecular cancer 2009;8:91 98. Kondo M, Ji L, Kamibayashi C, Tomizawa Y, Randle D, Sekido Y, et al. Overexpression of candidate tumor suppressor gene FUS1 isolated from the 3p21.3 homozygous deletion region leads to G1 arrest and growth inhibition of lung cancer cells. Oncogene 2001;20:6258-62 99. Pastuszak-Lewandoska D, Kordiak J, Migdalska-Sek M, Czarnecka KH, Antczak A, Gorski P, et al. Quantitative analysis of mRNA expression levels and DNA methylation profiles of three neighboring genes: FUS1, NPRL2/G21 and RASSF1A in non-small cell lung cancer patients. Respiratory research 2015;16:76 100. Demokan S, Chuang AY, Chang X, Khan T, Smith IM, Pattani KM, et al. Identification of guanine nucleotide-binding protein gamma-7 as an epigenetically silenced gene in head and neck cancer by gene expression profiling. International journal of oncology 2013;42:1427-36

101. Du L, Schageman JJ, Subauste MC, Saber B, Hammond SM, Prudkin L, et al. miR-93, miR-98, and miR-197 regulate expression of tumor suppressor gene FUS1. Molecular cancer research : MCR 2009;7:1234-43 102. Lee DY, Deng Z, Wang CH, Yang BB. MicroRNA-378 promotes cell survival, tumor growth, and angiogenesis by targeting SuFu and Fus-1 expression. Proc Natl Acad Sci U S A 2007;104:20350-5 103. Prudkin L, Behrens C, Liu DD, Zhou X, Ozburn NC, Bekele BN, et al. Loss and reduction of FUS1 protein expression is a frequent phenomenon in the pathogenesis of lung cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 2008;14:41-7 104. Uno F, Sasaki J, Nishizaki M, Carboni G, Xu K, Atkinson EN, et al. Myristoylation of the fus1 protein is required for tumor suppression in human lung cancer cells. Cancer Res 2004;64:2969-76 105. Xin J, Zhang XK, Xin DY, Li XF, Sun DK, Ma YY, et al. FUS1 acts as a tumor- suppressor gene by upregulating miR-197 in human glioblastoma. Oncology reports 2015;34:868-76 106. Li G, Kawashima H, Ji L, Ogose A, Ariizumi T, Umezu H, et al. Frequent absence of tumor suppressor FUS1 protein expression in human bone and soft tissue sarcomas. Anticancer research 2011;31:11-21 107. Rimkus T, Sirkisoon S, Harrison A, Lo HW. Tumor suppressor candidate 2 (TUSC2, FUS-1) and human cancers. Discovery medicine 2017;23:325-30 108. Li L, Yu C, Ren J, Ye S, Ou W, Wang Y, et al. Synergistic effects of eukaryotic coexpression plasmid carrying LKB1 and FUS1 genes on lung cancer in vitro and in vivo. Journal of cancer research and clinical oncology 2014;140:895-907 109. Ivanova AV, Ivanov SV, Pascal V, Lumsden JM, Ward JM, Morris N, et al. Autoimmunity, spontaneous tumourigenesis, and IL-15 insufficiency in mice with a targeted disruption of the tumour suppressor gene Fus1. The Journal of pathology 2007;211:591-601 110. Ito I, Ji L, Tanaka F, Saito Y, Gopalan B, Branch CD, et al. Liposomal vector mediated delivery of the 3p FUS1 gene demonstrates potent antitumor activity against human lung cancer in vivo. Cancer gene therapy 2004;11:733-9 111. Lu C, Stewart DJ, Lee JJ, Ji L, Ramesh R, Jayachandran G, et al. Phase I clinical trial of systemically administered TUSC2(FUS1)-nanoparticles mediating functional gene transfer in humans. PLoS One 2012;7:e34833 112. Meraz IM, Majidi M, Cao X, Lin H, Li L, Wang J, et al. TUSC2 Immunogene Therapy Synergizes with Anti-PD-1 through Enhanced Proliferation and Infiltration of Natural Killer Cells in Syngeneic Kras-Mutant Mouse Lung Cancer Models. Cancer immunology research 2018;6:163-77 113. Lin J, Sun T, Ji L, Deng W, Roth J, Minna J, et al. Oncogenic activation of c-Abl in non-small cell lung cancer cells lacking FUS1 expression: inhibition of c-Abl by the tumor suppressor gene product Fus1. Oncogene 2007;26:6989-96 114. Dai B, Yan S, Lara-Guerra H, Kawashima H, Sakai R, Jayachandran G, et al. Exogenous Restoration of TUSC2 Expression Induces Responsiveness to Erlotinib in Wildtype Epidermal Growth Factor Receptor (EGFR) Lung Cancer Cells through Context Specific Pathways Resulting in Enhanced Therapeutic Efficacy. PLoS One 2015;10:e0123967 115. Hood MI, Uzhachenko R, Boyd K, Skaar EP, Ivanova AV. Loss of mitochondrial protein Fus1 augments host resistance to Acinetobacter baumannii infection. Infection and immunity 2013;81:4461-9

116. Ji L, Roth JA. Tumor suppressor FUS1 signaling pathway. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2008;3:327-30 117. Meng J, Majidi M, Fang B, Ji L, Bekele BN, Minna JD, et al. The tumor suppressor gene TUSC2 (FUS1) sensitizes NSCLC to the AKT inhibitor MK2206 in LKB1-dependent manner. PLoS One 2013;8:e77067 118. Uzhachenko R, Ivanov SV, Yarbrough WG, Shanker A, Medzhitov R, Ivanova AV. Fus1/Tusc2 is a novel regulator of mitochondrial calcium handling, Ca2+- coupled mitochondrial processes, and Ca2+-dependent NFAT and NF-kappaB pathways in CD4+ T cells. Antioxidants & redox signaling 2014;20:1533-47 119. Coronas-Samano G, Baker KL, Tan WJ, Ivanova AV, Verhagen JV. Fus1 KO Mouse As a Model of Oxidative Stress-Mediated Sporadic Alzheimer's Disease: Circadian Disruption and Long-Term Spatial and Olfactory Memory Impairments. Frontiers in aging neuroscience 2016;8:268 120. Tan WJT, Song L, Graham M, Schettino A, Navaratnam D, Yarbrough WG, et al. Novel Role of the Mitochondrial Protein Fus1 in Protection from Premature Hearing Loss via Regulation of Oxidative Stress and Nutrient and Energy Sensing Pathways in the Inner Ear. Antioxidants & redox signaling 2017;27:489- 509 121. Uzhachenko R, Boyd K, Olivares-Villagomez D, Zhu Y, Goodwin JS, Rana T, et al. Mitochondrial protein Fus1/Tusc2 in premature aging and age-related pathologies: critical roles of calcium and energy homeostasis. Aging 2017;9:627- 49 122. Uzhachenko R, Issaeva N, Boyd K, Ivanov SV, Carbone DP, Ivanova AV. Tumour suppressor Fus1 provides a molecular link between inflammatory response and mitochondrial homeostasis. The Journal of pathology 2012;227:456-69 123. Deng WG, Kawashima H, Wu G, Jayachandran G, Xu K, Minna JD, et al. Synergistic tumor suppression by coexpression of FUS1 and p53 is associated with down-regulation of murine double minute-2 and activation of the apoptotic protease-activating factor 1-dependent apoptotic pathway in human non-small cell lung cancer cells. Cancer Res 2007;67:709-17 124. Deng WG, Wu G, Ueda K, Xu K, Roth JA, Ji L. Enhancement of antitumor activity of cisplatin in human lung cancer cells by tumor suppressor FUS1. Cancer gene therapy 2008;15:29-39 125. Orlandella FM, Di Maro G, Ugolini C, Basolo F, Salvatore G. TWIST1/miR- 584/TUSC2 pathway induces resistance to apoptosis in thyroid cancer cells. Oncotarget 2016;7:70575-88 126. Ren J, Yu C, Wu S, Peng F, Jiang Q, Zhang X, et al. Cationic liposome mediated delivery of FUS1 and hIL-12 coexpression plasmid demonstrates enhanced activity against human lung cancer. Current cancer drug targets 2014;14:167-80 127. Zhang B, Xu X, Qi Z, Peng L, Qiu B, Huo X. The FUS1 gene inhibits EC109 cell growth mediated by a lentivirus vector. British journal of biomedical science 2013;70:22-6 128. Xiaobo C, Majidi M, Feng M, Shao R, Wang J, Zhao Y, et al. TUSC2(FUS1)- erlotinib Induced Vulnerabilities in Epidermal Growth Factor Receptor(EGFR) Wildtype Non-small Cell Lung Cancer(NSCLC) Targeted by the Repurposed Drug Auranofin. Scientific reports 2016;6:35741 129. Yazlovitskaya EM, Uzhachenko R, Voziyan PA, Yarbrough WG, Ivanova AV. A novel radioprotective function for the mitochondrial tumor suppressor protein Fus1. Cell death & disease 2013;4:e687

130. Yazlovitskaya EM, Voziyan PA, Manavalan T, Yarbrough WG, Ivanova AV. Cellular oxidative stress response mediates radiosensitivity in Fus1-deficient mice. Cell death & disease 2015;6:e1652

CHAPTER II

Truncated glioma-associated oncogene homolog 1 (tGLI1) mediates mesenchymal

glioblastoma via transcriptional activation of CD44

Tadas K. Rimkus1,+, Richard L. Carpenter1,+,#, Sherona Sirkisoon1, Dongqin Zhu1, Boris

C. Pasche1,5, Michael D. Chan2,5,6, Glenn J. Lesser3,5,6, Stephen B. Tatter4,5,6, Kounosuke

Watabe1,5,6, Waldemar Debinski1,5,6, Hui-Wen Lo1,5,6

1 Department of Cancer Biology, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

2 Department of Radiation Oncology, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

3 Department of Hematology and Oncology, Wake Forest School of Medicine, Medical

Center Boulevard, Winston Salem, NC, 27157, USA;

4 Department of Neurosurgery, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

5 Comprehensive Cancer Center, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

6 Brain Tumor Center of Excellence, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA

# Current Address: Department of Biochemistry and Molecular Biology, Indiana

University School of Medicine-Bloomington, JH 308 1001 E. 3rd St., Bloomington, IN

47405

The manuscript was written through contributions of all authors. T.K.R., R.L.C., S.S., and

D.Z. developed the methodology and carried out experiments. B.C.P., M.D.C., G.J.L.,

S.B.T., K.W., and W.D. provided administrative and material support and review of the manuscript. T.K.R., R.L.C., and H.L. performed data analysis and wrote the manuscript.

R.L.C. and H.L. conceived the research and developed the methodology. All authors reviewed and have given approval to the final version of the manuscript. +These authors contributed equally.

Cancer Res., 2018, 78(10), pp 2589-2600.

DOI: 10.1158/0008-5472.CAN-17-2933

ABSTRACT

Molecular pathways driving mesenchymal glioblastoma (GBM) are still not well understood. We report in this study that truncated glioma-associated oncogene homolog

1, tGLI1, is a tumor-specific transcription factor that facilitates GBM growth, enriches in the mesenchymal subtype of GBM and glioma stem cells (GSCs), and promotes mesenchymal GSCs via transcriptionally activating CD44 expression. Using an orthotopic GBM xenograft mouse model, we observed that tGLI1-overexpressing tumors grew more aggressively with increased proliferation and angiogenesis compared to control and GLI1-overexpressing xenografts. Immunohistochemical analysis using tGLI1- and GLI1-selective antibodies we developed showed that tGLI1 was significantly expressed in GBM specimens but undetectable in normal brains whereas GLI1 was expressed in both tissues. Datamining further showed that tGLI1 activation signature

(tGAS) was correlated with glioma grade, tumor angiogenesis, and poor overall survival, and that GBMs with high tGAS were enriched with mesenchymal GBM and GSC gene signatures. Neurospheres contained increased levels of tGLI1, but not GLI1, compared to the monolayer culture. Mesenchymal GSCs expressed more tGLI1 than proneural

GSCs. Ectopic tGLI1 expression significantly enhanced the ability of GSCs to yield neurospheres in vitro and form aggressive tumors in mouse brains. Finally, we examined the ability of tGLI1 to regulate stemness genes and found that tGLI1 binds to and transactivates the promoter of the CD44 gene, a marker and mediator for mesenchymal

GSCs, leading to its expression. Collectively, these findings advanced our understanding of GBM biology by establishing tGLI1 as a novel transcriptional activator of CD44 and a novel mediator of mesenchymal GBM and GSCs.

INTRODUCTION

Glioblastoma (GBM) is the most common and most lethal brain tumor in adults accounting for 15% of all brain tumors. GBM Prognosis remains poor, with a median survival of 14 months and less than 5% of patients surviving five years post diagnosis

(1,2). Extensive genomic analyses have divided GBMs into 3-4 distinct subtypes with slight variations (3,4). The proneural and mesenchymal subtypes of GBM are commonly delineated across two classification systems. Mesenchymal GBMs tend to respond poorly to chemotherapy and radiation, associated with a worse prognosis (3-5).

Recent studies of GBM identified subpopulations of tumor cells exhibiting stem cell-like properties, including the ability to self-renew, persistently proliferate, and differentiate into multiple cell lineages (6). The subpopulation of stem-like cells, or glioma stem cells

(GSCs), is a key driver of tumor initiation, recurrence, and chemoresistance (7).

Recently, patient-derived GSC lines were identified to contain two distinct and mutually exclusive subtypes termed proneural and mesenchymal (8,9). CD44 was identified as a key marker of mesenchymal GSCs (8). High CD44 expression in GBM is predictive of poorer overall survival and is associated with increased GBM invasion, proliferation, and chemoresistance (10). Mesenchymal GSCs were found to be more aggressive and radioresistant; therefore, understanding the signaling pathways controlling the mesenchymal GSC phenotype is the key to developing targeted treatments for mesenchymal GSCs in GBM.

Sonic Hedgehog (SHH) pathway plays an essential role in GSC function (11). The glioma-associated oncogene homolog 1 (GLI1) family of zinc-finger transcription factors function as the terminal effectors of the SHH signaling pathway (12). Our laboratory discovered truncated GLI1 (tGLI1) as a novel alternatively spliced, gain-of-function variant of GLI1 with a 41 amino acid deletion spanning the entire exon 3 and part of

41 exon 4, but retains all of the known functional domains of GLI1 (13) . We reported that tGLI1 regulates known GLI1 target genes to a similar degree as wild-type GLI1 (13), but gained the ability to transcriptionally activate genes not regulated by GLI1, including

CD24, HPA1, TEM7, VEGF-A, VEGF-C, and VEGFR2, thus promoting cancer cell growth, migration, invasion, and angiogenesis (13-17). We also showed that tGLI1 is detectable only in cell lines, patient-derived xenografts, and primary specimens of GBM, but is undetectable in normal brain tissue or other normal human tissues (18). The tumor-specific expression pattern of tGLI1 in invasive breast cancer was reported in our previous study (19). Other groups have confirmed our findings that tGLI1 is detectable in triple-negative breast cancer (20) and malignant gliomas (21). Metastatic hepatoma cells also express tGLI1 (22).

GLI1 has been shown to drive transcription of pluripotency markers in cancer stem cells; however, the role that gain-of-function tGLI1 plays in cancer stem cells function remains unknown (11,23). It is also unknown whether tGLI1 promotes GBM progression using orthotopic xenograft mouse models. To address these gaps of knowledge, we initiated the current study including two different animal models, and found that tGLI1 drives the formation of larger, more proliferative, and more highly vascularized tumors than GLI1 in orthotopic GBM and GSC xenograft mouse models. Mechanistic studies provided evidence linking tGLI1 to mesenchymal GBM and GSCs. Our results also established tGLI1 as a novel transcriptional activator of the CD44 gene, a known marker and regulator of mesenchymal GSCs and also other cancer types.

MATERIALS AND METHODS

Cell lines and patient specimens

U373MG cells were obtained from ATCC and cultured according to their recommendations. Luciferase-expressing G48LL2 cells were developed by Dr.

Waldemar Debinski (24). BTCOE 4810/4795 cell lines were developed and validated from patient tumors at Brain Tumor Center of Excellence (25). Patient-derived Glioma

Stem Cells (GSCs) were kind gifts from Drs. Erik Sulman and Krishna Bhat at University of Texas MD Anderson Cancer Center (8). GSCs were passaged under sphere-forming conditions in serum-free DMEM/F12 growth medium supplemented with B27, FGF, and

EGF. Normal brain tissue microarray (BNC17011) and glioma tissue microarray

(GL2083) were purchased from US Biomax. Additional GBM samples were from Wake

Forest Brain Tumor Center of Excellence (25).

Generation of lentiviral GLI1- and tGLI1-expressing vectors and isogenic stable cell lines

The open reading frame for tGLI1 and GLI1 were cloned into the lentiviral expressing vector pCDH-CMV-MCS-EF1-RFP-Puro (CD516B-2) by System Biosciences (Palo Alto,

CA). Plasmids were sequenced to confirm insertion of tGLI1 and GLI1 open reading frames. Lentiviral packaging was performed using a 3rd generation lentiviral pPACK packaging system from System Biosciences. The pPACK packaging plasmids along with the tGLI1 or GLI1 lentiviral expressing plasmid were co-transfected into 293TN cells for

48-96 hrs and the cell culture medium was subsequently collected. Viral particles were concentrated, titers measured, and used to infect G48LL2 and GSC-28 cells. Infected cells were treated with puromycin and FACS sorted for RFP-positive cells.

Animal experiments

Female nude mice 6-8 weeks of age were used. Actively growing G48LL2 or GSC-28 isogenic cells were injected at a concentration of 1-5 x 105 cells in 5 µL PBS into the right frontal lobe. Mice were anesthetized with a ketamine/xylazine mixture to the coronal suture through a scalp incision. For bioluminescent imaging, xenograft-bearing mice were injected with d-luciferin intraperitoneally at 100 mg/kg body weight and then imaged weekly using PerkinElmer IVIS100 imager.

tGLI1 activated signature (tGAS) and Gene Set Enrichment Analysis (GSEA) tGAS was generated by averaging the median-centered expression of the six tGLI1 target genes (CD24, VEGFA, VEGFC, HPA1, TEM7, and VEGFR2) (13-17). GSEA was performed by generating the Gene Matrix file (.gmx) by using published signatures for

AngioMatrix (26), the TCGA GBM subtypes (4), the Phillips GBM subtypes (3), and the

GSC subtypes (8). The Gene Cluster Text file (.gct) was generated from the TCGA GBM dataset. The Categorical Class file (.cls) was generated based on the tGAS score or the

GLI1 expression of patients in the TCGA GBM dataset. The number of permutations for

GSEA was set to 1,000 and we used the TCGA gene list as the chip platform. For generation of heat maps, patients were divided into high or low tGAS score and the genes included in the map were genes within the published signatures for the indicated

GBM subtypes (3,4). Heat maps were generated using Morpheus software developed by the Broad Institute.

Locked Nucleic Acid Oligonucleotides tGLI1-specific Locked Nucleic Acid (LNA) oligonucleotides were custom made by

Exiqon. U373MG cells were transfected for forty-eight hours using Lipofectamine 2000

(Invitrogen) and then seeded for a neurosphere formation assay. tGLI1 mRNA expression was determined by real-time quantitative PCR. Sequence for control LNA oligonucleotides: C+T+G+T+C*T*T*C*A*G*T*T*C+A+A+C. Sequence for tGLI1-specific LNA

44 oligonucleotides: C+A+A+C+T*T*G*A*C*T*T*C*T+G+T+C. * indicates phosphorothioated base and + indicates LNA base.

Statistical analyses

Data are presented as mean±SE. The student’s t-test, Pearson Correlation, univariate/multivariate COX proportional hazards tests, and One-way ANOVA were performed using Sigma Plot version 11.0.

RESULTS tGLI1 promotes growth of GBM in an orthotopic mouse model.

Whether tGLI1 plays a role in promoting GBM intracranial growth has not been investigated. Herein, we generated isogenic cell lines from a low-passage G48LL2 human GBM cell line stably expressing empty vector, GLI1, or tGLI1 (Fig. 1A).

Expression levels for GLI1 and tGLI1 in these cell lines are similar to those found in

GBM specimens (Fig. 1A-bottom). The isogenic lines were implanted intracranially into female nude mice and tumor growth was tracked by weekly bioluminescence imaging.

Results showed that GBM cells expression tGLI1 formed larger tumors (Fig. 1B).

Representative images are shown in Figs. 1C-D. Analysis of brain sections using IHC indicated increased proliferation index (Ki-67 IHC) and increased microvessel density

(CD31 IHC) in tGLI1-expressing tumors (Figs. 1E-H). Representative IHC images are shown in Fig. 1I. Using custom GLI1- and tGLI1-specific antibodies that we developed and validated (17), we further examined GBM specimens and normal brain tissues for

GLI1 and tGLI1 expression, and found tGLI1 to be highly expressed in GBM samples, but undetectable in normal brain whereas GLI1 is expressed in both samples (Figs. 1J-

K). Results in Figure 1 demonstrate that tGLI1 promotes GBM tumor growth in the brain microenvironment and that tGLI1 is expressed in GBM but not in normal brain.

High tGLI1 activity is associated with poor overall survival of GBM patients and increased angiogenesis of GBM.

We asked whether tGLI1 activity can be used as a prognostic indicator for GBM. To indicate tGLI1 activation status, we created a tGLI1 activation signature (tGAS) using expression levels of its six target genes (13-17). We then analyzed a GEO dataset

(GSE4290) consisting of mRNA expression profiles of 23 normal brains, 45 grade II

46 gliomas, 31 grade III gliomas, and 81 grade IV gliomas (GBMs) for tGAS, and found

GBMs to have the highest tGAS scores (Fig. 2A). In contrast, GLI1 levels were not significantly different among 4 groups (Fig. 2B). Next, we examined whether tGAS was associated with clinical outcome for GBM patients using a TCGA GBM database and univariate/multivariate analyses, and observed that high tGAS resulted in high Hazard

Ratios (HRs) (Fig. 2C), suggesting that tGLI1 activation is independently associated with poor overall survival of GBM patients. Kaplan-Meier analysis further showed that GBM patients with high tGAS were associated with worse overall survival compared to patients with low tGAS (Fig. 2D). In contrast, GLI1 mRNA was not associated with overall survival (Fig. 2E).

Results in Fig. 1H indicated that tGLI1 promoted GBM vascularity. To confirm this observation, we correlated tGAS with vascularity markers, and found tGAS but not GLI1 to significantly associate with CD31 and VE-Cadherin (Figs. 2F-I). GSEA with the

AngioMatrix signature that has been shown to be associated with tumor vascularity in

GBM (26) further indicated that GBMs with high tGAS, but not high GLI1, had significant enrichment with the AngioMatrix signature (Figs. 2J-M). Results in Figure 2 demonstrate that tGLI1, but not GLI1, is associated with poor overall survival and enhanced tumor angiogenesis in GBM patients. tGLI1 activation is enriched in the mesenchymal subtype of GBM.

We examined whether the extent of tGLI1 activation differs among GBM subtypes. First, we divided the GBM cohort (TCGA) into two groups with high tGAS or low tGAS. We then used GSEA to determine the degrees of enrichment with established gene signatures for the four TCGA GBM subtypes, namely, mesenchymal, proneural, neural, and classical subtypes (4). As shown in Fig. 3A, GBM tumors with high tGAS were

47 enriched for the TCGA mesenchymal gene signature, but not for the signatures for the other three subtypes. Interestingly, GBM tumors with low tGAS were significantly enriched for the TCGA classical signature, which is less aggressive than the mesenchymal subtype. Furthermore, we analyzed the same GBM cohort for the enrichment for another set of GBM subtype-specific gene signatures, namely, Phillips mesenchymal, proneural, and proliferation signatures (3). Consistent with the results of

Fig. 3A, GBMs with high tGAS were enriched for the Phillips mesenchymal gene signature, but not the Phillips proneural or proliferation signatures (Fig. 3B).

Mesenchymal GBMs had the highest tGAS scores among the four TCGA subtypes (Fig.

3C). In contrast, high GLI1 expression was not associated with either TCGA or Phillips mesenchymal subtype, but associated with the TCGA classical subtype (Fig. 3D). GBMs with low GLI1 were enriched with both TCGA proneural and Phillips proneural gene signatures (Fig. 3E). These findings indicate that tGLI1 activation is a hallmark of the mesenchymal GBM. tGLI1 promotes neurosphere formation and transcriptionally activates CD44 expression.

Mesenchymal GBM is associated with poor patient survival and multi-drug resistance, key clinical aspects of GBM thought to be driven by the cancer stem cell subpopulation

(3,5,27). The role of GLI1 in GSC function has been studied; however, the role that tGLI1 plays in GSCs is unknown. We first compare monolayer culture with GSC- containing neurospheres for tGLI1 and GLI1 expression, and found tGLI1 but not GLI1 to be enriched in neurospheres (Fig. 4A). We further found that tGLI1-overexpressing cells formed significantly more neurospheres (Figs. 4B-C). Additionally, we found that knocking down tGLI1 using tGLI1-specific Locked Nucleic Acid (LNA) oligonucleotides inhibited neurosphere formation in GBM cells (Fig. 4D). To identify the mechanism by

48 which tGLI1 promotes a GBM stem-cell phenotype, we determined the effects of tGLI1 on known stem cell-related genes and found that tGLI1-overexpressing GBM cells showed increased expression of CD44, along with a decrease in Sox2, Nanog, OCT4, and CD133 expression (Fig. 4E). Flow cytometry confirmed increased CD44(+) cells in tGLI1-expressing GBM cells but no significant change in the CD133(+) population (Fig.

4F). CD44 is regarded as the marker for the mesenchymal GSCs while CD133 is the marker for the proneural GSCs (8). Using the ChIP assay, we detected preferential binding of tGLI1, but not GLI1, to three regions of the CD44 gene promoter (Fig. 4G;

Supplemental Fig. II). Using the luciferase assay, we showed that tGLI1 transactivated the CD44 promoter in two GBM cell lines and HEK293 cells (Fig. 4H) and that SHH enhanced tGLI1-mediated activation of the CD44 gene promoter (Fig. 4I). Results in

Figure 4 indicate a novel important role that tGLI1 plays in GBM stem-like cells and in the transcriptional activation of CD44. tGLI1 is preferentially expressed and activated in the mesenchymal subtype of

GSC.

Recent studies classified GSCs isolated from GBM specimens into the proneural (PN) or mesenchymal (MES) subtypes based on gene expression profiles (8,9). CD44 was defined as the marker for the MES GSCs whereas CD133 was for the PN GSCs. Since tGLI1 binds to and transactivates the CD44 promoter and tGLI1 is preferentially activated in MES GBM, we determined whether tGLI1 expression is associated with the

MES subtype of GSC. Here, we examined patient-derived GSCs, two MES and two PN

GSC lines, and found the MES GSCs to express higher levels of tGLI1 and lower levels of GLI1 compared to PN GSCs (Fig. 5A). We also observed that MES GSC lines had higher tGAS than the PN GSC lines (Fig. 5B). We further analyzed a TCGA GBM dataset for the relationship between tGAS and GSC signatures, and observed that

GBMs with high tGAS are enriched for the MES GSC signature, but not the PN GSC signature (Fig. 5C-left). In contrast, GLI1 was not enriched for either GSC signature

(Fig. 5C-right).

We found that tGLI1 ectopic expression enhanced the neurosphere-forming capability of both PN and MES GSC lines (Figs. 5D-E), and led to an increase in CD44 expression and a decrease in CD133 expression at both the mRNA and protein levels (Figs. 5F-G).

Further analysis of the four GSC lines indicated a pattern where high tGLI1-expressing

MES GSCs expressed higher levels of CD44 and lower levels of CD133, compared to low tGLI1-expressing PN GSCs (Fig. 5H). Consistent with these observations, analysis of the TCGA GBM dataset revealed a positive correlation between tGAS and CD44, but not tGAS and CD133 (Fig. 5I). Collectively, results in Figure 5 demonstrate that tGLI1 is preferentially expressed and activated in the MES GSC over PN GSC, enhances expression of CD44, a MES GSC marker, and positively correlates with CD44 expression in a GBM cohort.

Increased tGLI1 expression enhanced the propensity of GSC to form xenografts.

We have shown that tGLI1 ectopic expression promoted GBM growth in an orthotopic mouse model and that tGLI1 made GSCs formed greater numbers of neurospheres. In light of these observations, we aimed to validate our GSC findings in vivo. We first generated three isogenic GSC-28 cell lines with stable expression of control, GLI1 or tGLI1 lentiviral vectors (Fig. 6A). Of note, expression levels for GLI1 and tGLI1 in these cell lines are similar to those found in GBM specimens as shown by IHC (Fig. 6A- bottom). These lines were then implanted into the mouse brains; mice were imaged weekly. As shown in the growth curves and representative tumor images in Figs. 6B-D, increased tGLI1 expression rendered GSC-28 cells more aggressive in growths. Mice

50 bearing tGLI1-expressing GSC-28 xenografts had a shortened survival time (Fig. 6E).

IHC analyses further demonstrated that tGLI1-expressing GSC-28 tumors had the highest proliferative index (Ki-67 IHC) and microvessel density (CD31 IHC) (Figs. 6F-J).

In summary, these results indicate that tGLI1 renders GSCs more aggressive in growth and angiogenesis in the brain microenvironment.

DISCUSSION

We made the following important novel observations in this study: a) tGLI1 promotes neurosphere-forming ability of GBM and GSCs in vitro, and their intracranial growth and angiogenesis in vivo; b) tGLI1 is highly expressed in GBM but undetectable in normal brain whereas GLI1 is equally expressed in both tissues ; c) tGLI1 is predominantly activated in the mesenchymal subtype of GBM and GSCs that are more aggressive among different subtypes (4).; d) GBM patients with high tGLI1 activity in their tumors had shortened overall survival of GBM patients and increased tumor angiogenesis, compared to those with low tGLI1 activity; and e) tGLI1 functions as a transcriptional activator of CD44. By reporting these findings, our study advances the biological understanding of GBM and GSCs and transcriptional regulation of an important stem cell marker/mediator CD44.

Our bioinformatics analyses revealed tGLI1 activation as a hallmark of the mesenchymal

GBM and GSCs, which is an important finding. We used two independent mesenchymal gene signatures and found both signatures to be highly enriched in GBMs and GSCs with high tGAS, an indicator for tGLI1 transcriptional activity. Of note, the classification of

GBM into distinct subtypes was first reported by Phillips et al. in 2006 with three distinct subtypes (3), while a more recently study in 2010 by Verhaak et al. classified GBM into four distinct subtypes (4). Both studies identified a common subtype named the mesenchymal subtype according to the expression of genes associated with a mesenchymal cell phenotype. Notably, several earlier studies identified mediators and markers for the mesenchymal subtype of GBM and GSCs. For example, RTVP-1 was found to express at a higher level in mesenchymal GBM associated with tumor recurrence and poor clinical outcome (28). RTVP-1 overexpression induced mesenchymal differentiation of human neural stem cells, whereas silencing RTVP-1

52 inhibited the mesenchymal transformation and stemness of GSCs. An important future task is to explore the potential crosstalk between RTVP-1 and tGLI1 in regulating mesenchymal GSCs.

It is also important future task to investigate the potential interactions between tGLI1 and

TNF-a/NF-kB in light of the observations that both are enriched in the CD44(+) mesenchymal GSCs and that the proneural GSCs can undergo differentiation to a mesenchymal state in a TNF-a/NF-kB-dependent manner (8,29). Most recently, S100A4 was reported as a novel biomarker of GSCs that is enriched in cells with tumor-initiating and sphere-forming abilities; selective ablation of S100A4-expressing cells blocked tumor growth in vitro and in vivo (30). Whether tGLI1 crosstalks with S100A4, thereby promoting mesenchymal GSCs is unknown but this line of future research is warranted.

Mesenchymal GBM and GSCs are not only more tumorigenic but also more resistant to radiation therapy (8). Given the ability of tGLI1 to promote mesenchymal GBM and

GSCs, it is possible that tGLI1-expressing cells are more resistant to radiation therapy.

Although the direct role of tGLI1 in radiation resistance has not been reported, it has been shown that tumor cells with hyperactive SHH-GLI1 signaling are more resistant to radiation therapy (31-33). Since tGLI1 functions as a gain-of-function GLI1 and tGLI1 has a higher propensity than GLI1 to promote GSCs in vitro and in vivo, we speculate that tGLI1(+) GSCs are more resistant to radiation therapy compared to GLI1(+) GSCs, which could be tested in future studies.

CD44 is regarded as a marker for cancer stem cells for a number of cancers, including breast cancer, pancreatic cancer, and GBM (8,34,35) that is associated with tumor initiation and progression (34,36-38). However, emerging evidence suggests that CD44 contributes to the stem cell phenotype via various mechanisms, such as osteopontin

53 signaling and promotion of HIF-2α activity (39). However, transcriptional regulation of

CD44 gene is still not well understood. It has been reported that EGR1 transcription factor can induce CD44 expression antigen-stimulated B cells (40). NF-κB has been shown to upregulate CD44 expression in GBM (8) and breast cancer (41). NF-κB cooperates with AP-1 to bind to a cis-element of the CD44 promoter, leading to CD44 expression in breast cancer in a cell type-specific manner (42). Y-box binding protein-1 transcription factor induces CD44 expression in breast cancers (43). Surprisingly,

FOXO3 tumor suppressor has been shown to induce CD44 expression in pancreatic cancer (44). Our discovery that tGLI1 directly activates CD44 gene expression sheds important new light into the molecular mechanisms contributing to high CD44 expression in GSCs and possibly cancer stem cells in other tumor types.

Our present and previous studies uncovered that tGLI1 is expressed in a tumor-specific fashion in GBM and breast cancer (13,15-17,19). Expression of tGLI1 in malignant gliomas (21) and breast cancer (20) has been confirmed by other groups. Interestingly, a recent study detected tGLI1 in metastatic hepatoma cells (22). The mechanisms for the tumor-specific expression of tGLI1 are, however, still not elucidated. Since the splicing machinery is highly dysregulated in cancers (45), it is likely that the splicing factors that synthesize tGLI1 are aberrantly overexpressed in GBM, breast cancer, and metastatic hepatomas. Identification of these splicing factors constitutes an important task that could lead to strategies that inhibit tumor progression through inhibiting tGLI1 production.

Evidence from our laboratory and those of other groups suggests that tGLI1 may be regarded as novel therapeutic target for multiple cancer types (13,15-17,19,20,22,46).

This notion is supported by the observations that tGLI1 is expressed tumor specifically and that tGLI1 plays an important role in tumor growth, angiogenesis and cancer stem

54 cell renewal. We further speculate that targeting tGLI1, rather than GLI1, could minimize normal tissue toxicity because tGLI1 is only detected in cancer tissues whereas GLI1 is expressed in both cancerous and normal tissues. Smoothened inhibitors can inhibit both tGLI1 and GLI1; however, their efficacy is only modest in GBM and breast cancer due to their inability to suppress non-canonical smoothened-independent activation of tGLI1 and GLI1 (47). Our study provides novel insights into the biology of GBM and GSCs, particularly those belonging to the mesenchymal subtype, defines tGLI1 as a novel mediator of GBM growth and GSC self-renewal, and establishes tGLI1 as a novel transcriptional activator of CD44.

FIGURES

Figure 1. tGLI1 is expressed in a tumor-specific fashion and promotes intracranial GBM growth. A) Isogenic G48LL2 GBM cell lines carrying lentiviral mock, GLI1, or tGLI1 vector were subjected to immunoblotting for GLI1 and tGLI1 expression (top panel). Expression levels for GLI1 and tGLI1 in these cell lines are similar to those found in GBM specimens as shown by IHC (bottom panel). B) tGLI1 rendered GBM more aggressive in growth. Isogenic luciferase-expressing G48LL2 cell lines were injected into the right frontal lobe of female nude mice (N=5 per group) and tumor growth was assessed weekly via bioluminescent imaging. C) Representative bioluminescent images of actively growing tumors at Day 56. D) Representative bioluminescent images of ex vivo mouse brains. E- I) tGLI1-overexpressing GBM xenografts were more proliferative and more vascularized Mouse brains were subjected to H&E staining and IHC with indicated antibodies. Immunostained sections were scored by a pathologist and H-scores were calculated. Panel I shows representative IHC images. J-K) tGLI1 is highly expressed in GBM specimens but not in normal brain tissues. A cohort of normal healthy brain tissues (N=80) and GBM patient samples (N=63) were subjected to IHC using GLI1- and tGLI1- specific antibodies. Immunostained sections were scored by a pathologist to derive H- scores. Panel K shows representative IHC images. Student’s t-test was used to compute p-values.

Figure 2. High tGAS is associated with poor overall survival of GBM patients and increased angiogenesis of GBM samples. A-B) tGLI1 is most activated in GBM across four grades of gliomas. tGAS scores (A) and GLI1 mRNA expression (B) were determined for patient tumors in GSE4290 dataset across normal healthy brain and glioma samples. Student’s t-test was used to compute

58 p-values. NS, non-significant. C) Increased tGLI1 activity is independently associated with poor overall survival of GBM patients. Univariate and multivariate COX proportional hazards were calculated using tGAS score, GLI1 expression, patient age, patient sex, Karnofsky Performance Score (KPS), and tumor size as variables. HR, hazard ratio. The TCGA GBM dataset was used with the outcome variable being overall survival. D-E) Patients with high tGAS in their GBMs had worse overall survival compared to those to low tGAS. Kaplan-Meier survival graphs were drawn using high or low tGAS score (D) or GLI1 expression (E) and data from the TCGA GBM dataset. Log-rank method was used to compute p-values. F-I) tGAS positively correlated with GBM angiogenesis. tGAS score (F-G) or GLI1 expression (H-I) was correlated with markers of tumor vascularity using CD31 (F,H) and VE-cadherin (G,I) using regression analysis. J-K) GBMs with high tGAS were enriched with angiogenesis gene signature. GSEA was performed using the AngioMatrix signature that is representative of GBM angiogenesis. Patients were separated by high or low tGAS score (J) or GLI1 expression (K) using the TCGA GBM dataset. L-M) A positive correlation between tGAS and angiogenesis in GBM specimens. AngioMatrix signature score was correlated with tGAS score (L) or GLI1 expression (M) using the TCGA GBM dataset.

Figure 3. GBMs with high tGAS are enriched with the mesenchymal GBM gene signatures. The TCGA dataset with the gene expression profile of 165 GBMs was used. A-B) GBM tumors with high tGAS were enriched for the TCGA and Phillips mesenchymal gene

60 signatures, but not for those for the other subtypes. Heat maps were drawn using Morpheus software with patients separated by high or low tGAS score using genes in the signatures for each of the four TCGA GBM subtypes (A) or the three Phillips GBM subtypes (B). Right: GSEA. C) The mesenchymal subtype of GBMs had the highest tGAS scores among the four TCGA subtypes. tGAS score was determined for the four TCGA GBM subtypes. D-E) High GLI1 mRNA expression was not associated with either the TCGA or the Phillips mesenchymal subtype, but associated with the TCGA classical subtype. GSEA was performed using the signatures for the TCGA GBM subtypes (D) or the Phillips GBM subtypes (E) and patients were separated by high or low GLI1 mRNA expression.

Figure 4. tGLI1 promotes neurosphere formation and transcriptionally activates CD44 expression. A) tGLI1 was significantly enriched in the neurospheres compared to the monolayer GBM cells. G48LL2 and BTCOE 4810 cell lines were collected under monolayer and neurosphere-forming conditions. Total RNA from cells were subjected to qPCR for tGLI1 and GLI1 levels. B-C) tGLI1-overexpressing cells formed significantly more

62 neurospheres, suggesting an increase in the stem-like cell population. G48LL2 (B) and U373MG (C) cells with stable expression of vector, GLI1, or tGLI1 were subjected to the neurosphere assay. D) tGLI1 knockdown significantly inhibited neurosphere formation. U373MG cells transfected with tGLI1-specific Locked Nucleic Acid (LNA) oligonucleotides were subjected to the neurosphere assay. E) tGLI1 enhanced CD44 expression. Total RNA from G48a cells with stable expression of either vector or tGLI1 were subjected to qPCR for the indicated genes. F) tGLI1-expressing GBM cells had increased CD44(+) cells. Isogenic G48LL2 cell lines with stable expression of control or tGLI1 vector were subjected to flow cytometry with CD44 or CD133 antibodies. G) Preferential binding of tGLI1, but not GLI1, to the CD44 promoter. U373MG cells with transient expression of control vector or tGLI1 were subjected to the ChIP assay followed by PCR using primers for three regions of the CD44 gene promoter. H) tGLI1 transactivated the CD44 promoter in two GBM cell lines and HEK293 cells. Cell lines were transiently transfected with control or tGLI1 vector along with the CD44 promoter luciferase reporter, stimulated with SHH (100ng/mL) for 4 hrs, and subjected to the luciferase assay. I) SHH stimulation enhanced tGLI1-mediated activation of the CD44 gene promoter. BTCOE 4795 and U373MG cell lines were transiently transfected with control or tGLI1 vector together with the CD44 promoter luciferase reporter. Cells were then treated with or without SHH (100 ng/mL) for 4 hrs, harvested, and subjected to the luciferase assay. Student’s t-test was performed to calculate p-values. NS, non- significant. All experiments were done at least three times to derive means and standard deviations.

Figure 5. tGLI1 is preferentially expressed and activated in the mesenchymal subtype of GSC.

A) MES GSCs expressed higher levels of tGLI1 and lower levels of GLI1 compared to PN GSCs. Four GSC lines that were previously characterized as MES or PN subtype were subjected to total RNA extraction followed by RT-qPCR for tGLI1 and GLI1 expression levels. B) tGAS was significantly higher MES GSC lines compared to the PN GSC lines. tGAS scores were calculated for PN GSC (N=11) and MES GSC (N=6) lines that were isolated and profiled for expression by Bhat et al. (8). C) GBMs with high tGAS are enriched for the MES GSC signature, but not the PN GSC signature. The TCGA GBM dataset was analyzed by GSEA for the extent of enrichment with the MES and PN GSC signatures (8). Patients were divided into two groups according to tGAS (left) or GLI1 expression (right). D-E) tGLI1 overexpression increased neurosphere-forming capability of both PN (D) and MES (E) GSC lines. GSC-11 (PN) and GSC-28 (MES) transiently transfected with control vector or tGLI1 vector were subjected to the neurosphere assay (left) and RT-qPCR for tGLI1 expression levels (right). F-G) tGLI1 increased CD44 and decreased CD133 expression at the mRNA (F) and protein levels (G). H) High tGLI1-expressing MES GSCs expressed higher levels of CD44 and lower levels of CD133, compared to low tGLI1-expressing PN GSCs. Extracted total RNA was subjected to RT-qPCR for expression of tGLI1, CD44, and CD133. I) A positive correlation between tGAS and CD44, but not CD133 in GBM cohort (N=165). tGAS score was correlated with expression levels of CD44 or CD133 in the TCGA GBM dataset using Pearson correlation. Student’s t-test was used to calculate p-values.

Figure 6. Increased tGLI1 expression enhanced the propensity of GSC to form xenografts. A) Generation of isogenic GSC-28 cells with stable expression of control, GLI1, or tGLI1 vector. Cells were analyzed by immunoblotting for GLI1 and tGLI1 levels (top panel). Expression levels for GLI1 and tGLI1 in these cell lines are similar to those found in

GBM specimens as shown by IHC (bottom panel). B) GSC-28 cells expressing tGLI1 formed larger tumors compared to GLI1- or vector-expressing cells. Isogenic lines were implanted into the right frontal lobe of nude mice (N=6 per group) and tumor growth was assessed weekly via bioluminescent imaging. C) Representative images of actively growing tumors from animals. D) Representative images of ex vivo mouse brains. E) Mice bearing tGLI1-expressing GSC-28 xenografts had a shortened survival time. Kaplan-Meier survival graph is shown. Log rank test was used to determine p-values. F) Mouse brains were subjected to H&E staining and IHC with indicated antibodies. Representative images are shown. G-J) tGLI1-expressing GSC-28 tumors had the highest proliferative index and microvessel density. Immunostained mouse brains were scored to determine H-scores. Student’s t-test was used to compute p-values.

Supplemental Figure 1: CD44 Promoter Regions for tGLI1 ChIP. Three regions of the CD44 gene promoter, relative to the transcription start site (TSS), were assessed for tGLI1 binding by ChIP. Region 1 spans -962 bp to -696 bp, region 2 spans -723 bp to - 318 bp, and region 3 spans -357 bp to -17 bp.

TABLES

Supplemental Table I: Primers for RT-qPCR.

Gene Forward Primer (5’->3’) Reverse Primer (5’->3’) tGLI1 GTGTGGGGACAGAAGTCAA GTGCGGATAACCGTCTGC GLI1 CACCAAGCTAACCTCATGTC CGGGGAGAAGAAAAGAGTGGG CD44 TCAGAGGAGTAGGAGAGAGGAA AAGTCAAAGTAACAATAAGAGTGG AC TCA Sox2 GGAGTTGTCAAGGCAGAGAAGA GAGAGAGGCAAACTGGAATC G Nanog CTAAGAGGTGGCAGAAAAACA CTGGTGGTAGGAAGAGTAAAGG OCT4 TGGTCCGAGTGTGGTTCTGTAA TGTGCATAGTCGCTGCTTGAT CD133 ACCCATTGGCATTCTCTTTG TTTTGGATTCATATGCCTTGTGT GAPDH ACTGCCAACGTGTCAGTGG GTGTCGCTGTTGAAGTCAGA

Supplemental Table II: Primers for tGLI1 ChIP on CD44 Promoter.

CD44 Promoter Forward Primer (5’->3’) Reverse Primer (5’->3’) Region CD44 CAATCTCAAAAGGCTTCCCCT AACCATCCACCATCCTCTTCTC Region 1 CD44 GGGTGGAGAAGAGGATGGTGGA GCTTCTTGGCAGAACAGCTC Region 2 CD44 GGCCATCAGTAGCTTTCCCT GTCAGGACAGAGGATGACCG Region 3

SUPPLEMENTARY MATERIALS AND METHODS

Immunoblotting and Immunohistochemistry (IHC)

Immunoblotting and IHC were performed as we previously described (15,17). Antibodies for immunoblotting include GLI1 antibody (Cell Signaling Technology/CST; L42B10), a- tubulin (Sigma; B-512), CD44 (CST; 3578), CD133 (CST; D4W4N), OCT4 (CST; 9B7), and Sox2 (CST; L1D6A2). Antibodies used in IHC include Ki-67 (Neomarkers; RB-9043-

R7) and CD31 (BD Biosciences; 550274). GLI1- and tGLI1-specific rabbit polyclonal antibodies for IHC and immunoblotting were developed by our lab as previously described (15-17). Histologic scores (H-Scores) were computed from both % positivity

(A%, A=1-100) and intensity (B=0-3) using the equation, H-Score=A x B. CD31 IHC was used to mark tumor vasculature in which blood vessels in each of 8-10 microscopic fields were counted under a 40X objective using a microscope to derive microvessel density, via a standard method and the formula, # of vessels/mm2 (48).

RNA extraction and quantitative PCR

Total RNA was isolated using Promega SV Total RNA Isolation kit and subjected to quantitative PCR as we previously described (15,17). Primers for qPCR are described in

Supplemental Table I.

Flow cytometry

GBM cells were trypsinized and washed twice in PBS. These cells were then incubated with CD133-PE (Miltenyi Biotech) or CD44-APC (BioLegend) for 20 min. Percentage of cells positive for CD44 or CD133 was measured using a BD Accuri instrument.

Neurosphere assay

Adherent cells were seeded in neurosphere medium composed of base Neurobasal medium (LifeTechnologies) with 2% B27 (Sigma), 10 ng/mL FGF (Sigma), 100 ng/mL

EGF (Sigma), and 100 ng/mL sonic hedgehog (Sigma). Cells were seeded into 24-well

69 low-attachment plates (Corning) in neurosphere medium and spheres were counted 7-

10 days later.

Chromatin immunoprecipitation (ChIP) and promoter reporter assays

A ChIP Assay kit (Upstate/Millipore) was used as we previously described (17). A rabbit polyclonal GLI1/tGLI1 antibody (Santa Cruz; H-300) was used for the immunoprecipitation as we have described previously (17). Rabbit normal IgG (Sigma) served as negative immunoprecipitation controls and input chromatin was used as loading controls for PCR. The primers used for detection of the CD44 gene promoter are in Supplemental Table 2. Renilla luciferase expression vector, pRL-CMV, was used to control for transfection efficiency. Forty-eight hours after transfection, the cells were lysed and luciferase activity measured using the Firefly and Renilla Luciferase Assay Kit

(Biotium), as we previously described [17; 18]. Relative promoter activity was computed by normalizing the Firefly luciferase activity against that of the Renilla luciferase. The firefly luciferase plasmid under the control of the CD44 gene promoter, CD44P-pGL3, was purchased from Addgene (plasmid #19122) (49).

ACKNOWLEDGEMENTS

We thank Drs. Erik Sulman and Krishna Bhat at University of Texas MD Anderson

Cancer Center for gifting us the GSC cell lines. We also acknowledge the financial support from the NIH; R01-NS087169 (to HWL) and P30-CAO12197 (to BP).

REFERENCES

1. Kumthekar PU, Macrie BD, Singh SK, Kaur G, Chandler JP, Sejpal SV. A review of management strategies of malignant gliomas in the elderly population. American Journal of Cancer Research 2014;4:436-44 2. Ostrom QT, Gittleman H, Farah P, Ondracek A, Chen Y, Wolinsky Y, et al. CBTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in the United States in 2006-2010. Neuro-Oncology 2013;15:ii1-ii56 3. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 2006;9:157-73 4. Verhaak RGW, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, et al. An integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR and NF1. Cancer cell 2010;17:98 5. Colman H, Zhang L, Sulman EP, McDonald JM, Shooshtari NL, Rivera A, et al. A multigene predictor of outcome in glioblastoma. Neuro-Oncology 2010;12:49-57 6. Lathia JD, Mack SC, Mulkearns-Hubert EE, Valentim CLL, Rich JN. Cancer stem cells in glioblastoma. Genes & Development 2015;29:1203-17 7. Stopschinski BE, Beier CP, Beier D. Glioblastoma cancer stem cells--from concept to clinical application. Cancer letters 2013;338:32-40 8. Bhat KP, Balasubramaniyan V, Vaillant B, Ezhilarasan R, Hummelink K, Hollingsworth F, et al. Mesenchymal differentiation mediated by NF-kappaB promotes radiation resistance in glioblastoma. Cancer cell 2013;24:331-46 9. Mao P, Joshi K, Li J, Kim SH, Li P, Santana-Santos L, et al. Mesenchymal glioma stem cells are maintained by activated glycolytic metabolism involving aldehyde dehydrogenase 1A3. Proceedings of the National Academy of Sciences of the United States of America 2013;110:8644-9 10. Mooney KL, Choy W, Sidhu S, Pelargos P, Bui TT, Voth B, et al. The role of CD44 in glioblastoma multiforme. Journal of clinical neuroscience : official journal of the Neurosurgical Society of Australasia 2016;34:1-5 11. Clement V, Sanchez P, de Tribolet N, Radovanovic I, Ruiz i Altaba A. HEDGEHOG-GLI1 Signaling Regulates Human Glioma Growth, Cancer Stem Cell Self-Renewal, and Tumorigenicity. Current Biology 2007;17:165-72 12. Stecca B, Ruiz IAA. Context-dependent regulation of the GLI code in cancer by HEDGEHOG and non-HEDGEHOG signals. Journal of molecular cell biology 2010;2:84-95 13. Lo HW, Zhu H, Cao X, Aldrich A, Ali-Osman F. A novel splice variant of GLI1 that promotes glioblastoma cell migration and invasion. Cancer research 2009;69:6790-8 14. Cao X, Geradts J, Dewhirst M, Lo H-W. Upregulation of VEGF-A and CD24 Gene Expression by the tGLI1 Transcription Factor Contributes to the Aggressive Behavior of Breast Cancer Cells. Oncogene 2012;31:104-15 15. Carpenter RL, Paw I, Zhu H, Sirkisoon S, Xing F, Watabe K, et al. The gain-of- function GLI1 transcription factor TGLI1 enhances expression of VEGF-C and TEM7 to promote glioblastoma angiogenesis. Oncotarget 2015;6:22653-65 16. Han W, Carpenter RL, Lo H-W. TGLI1 Upregulates Expression of VEGFR2 and VEGF-A, Leading to a Robust VEGF-VEGFR2 Autocrine Loop and Cancer Cell Growth. Cancer Hallmarks 2013;1:28-37

17. Zhu H, Carpenter RL, Han W, Lo H-W. The GLI1 Splice Variant TGLI1 Promotes Glioblastoma Angiogenesis and Growth. Cancer letters 2014;343:10.1016/j.canlet.2013.09.014 18. Carpenter RL, Lo HW. Hedgehog pathway and GLI1 isoforms in human cancer. Discovery medicine 2012;13:105-13 19. Cao X, Geradts J, Dewhirst MW, Lo HW. Upregulation of VEGF-A and CD24 gene expression by the tGLI1 transcription factor contributes to the aggressive behavior of breast cancer cells. Oncogene 2012;31:104-15 20. Di Mauro C, Rosa R, D'Amato V, Ciciola P, Servetto A, Marciano R, et al. Hedgehog signalling pathway orchestrates angiogenesis in triple-negative breast cancers. British journal of cancer 2017;116:1425-35 21. Dovydas Gricius, Kazlauskas A. Analysis of oncogene GLI1 protein expression levels between differing grades of astrocytoma. BIOLOGIJA 2014;60:142-7 22. Fan YH, Ding J, Nguyen S, Liu XJ, Xu G, Zhou HY, et al. Aberrant hedgehog signaling is responsible for the highly invasive behavior of a subpopulation of hepatoma cells. Oncogene 2015 23. Ruiz i Altaba A. Hedgehog signaling and the Gli code in stem cells, cancer, and metastases. Science signaling 2011;4:pt9 24. Debinski W, Gibo DM. Fos-Related Antigen 1 Modulates Malignant Features of Glioma Cells. Molecular Cancer Research 2005;3:237-49 25. Ferluga S, Tomé CML, Herpai DM, D'Agostino R, Debinski W. Simultaneous targeting of Eph receptors in glioblastoma. Oncotarget 2016;7:59860-76 26. Langlois B, Saupe F, Rupp T, Arnold C, van der Heyden M, Orend G, et al. AngioMatrix, a signature of the tumor angiogenic switch-specific matrisome, correlates with poor prognosis for glioma and colorectal cancer patients. Oncotarget 2014;5:10529-45 27. Pelloski CE, Mahajan A, Maor M, Chang EL, Woo S, Gilbert M, et al. YKL-40 expression is associated with poorer response to radiation and shorter overall survival in glioblastoma. Clinical cancer research : an official journal of the American Association for Cancer Research 2005;11:3326-34 28. Giladi ND, Ziv-Av A, Lee HK, Finniss S, Cazacu S, Xiang C, et al. RTVP-1 promotes mesenchymal transformation of glioma via a STAT-3/IL-6-dependent positive feedback loop. Oncotarget 2015;6:22680-97 29. Kim SH, Ezhilarasan R, Phillips E, Gallego-Perez D, Sparks A, Taylor D, et al. Serine/Threonine Kinase MLK4 Determines Mesenchymal Identity in Glioma Stem Cells in an NF-kappaB-dependent Manner. Cancer Cell 2016;29:201-13 30. Chow KH, Park HJ, George J, Yamamoto K, Gallup AD, Graber JH, et al. S100A4 is a biomarker and regulator of glioma stem cells that is critical for mesenchymal transition in glioblastoma. Cancer Res 2017 31. Zhou J, Wu K, Gao D, Zhu G, Wu D, Wang X, et al. Reciprocal regulation of hypoxia-inducible factor 2alpha and GLI1 expression associated with the radioresistance of renal cell carcinoma. International journal of radiation oncology, biology, physics 2014;90:942-51 32. Gan GN, Eagles J, Keysar SB, Wang G, Glogowska MJ, Altunbas C, et al. Hedgehog signaling drives radioresistance and stroma-driven tumor repopulation in head and neck squamous cancers. Cancer Res 2014;74:7024-36 33. Chen YJ, Lin CP, Hsu ML, Shieh HR, Chao NK, Chao KS. Sonic hedgehog signaling protects human hepatocellular carcinoma cells against ionizing radiation in an autocrine manner. International journal of radiation oncology, biology, physics 2011;80:851-9

34. Jijiwa M, Demir H, Gupta S, Leung C, Joshi K, Orozco N, et al. CD44v6 Regulates Growth of Brain Tumor Stem Cells Partially through the AKT-Mediated Pathway. PLOS ONE 2011;6:e24217 35. Li C, Heidt DG, Dalerba P, Burant CF, Zhang L, Adsay V, et al. Identification of pancreatic cancer stem cells. Cancer research 2007;67:1030-7 36. Bourguignon LYW, Peyrollier K, Xia W, Gilad E. Hyaluronan-CD44 Interaction Activates Stem Cell Marker Nanog, Stat-3-mediated MDR1 Gene Expression, and Ankyrin-regulated Multidrug Efflux in Breast and Ovarian Tumor Cells. Journal of Biological Chemistry 2008;283:17635-51 37. Breyer R, Hussein S, Radu DL, Putz KM, Gunia S, Hecker H, et al. Disruption of intracerebral progression of C6 rat glioblastoma by in vivo treatment with anti- CD44 monoclonal antibody. Journal of neurosurgery 2000;92:140-9 38. Prince ME, Sivanandan R, Kaczorowski A, Wolf GT, Kaplan MJ, Dalerba P, et al. Identification of a subpopulation of cells with cancer stem cell properties in head and neck squamous cell carcinoma. Proceedings of the National Academy of Sciences 2007;104:973-8 39. Pietras A, Katz AM, Ekstrom EJ, Wee B, Halliday JJ, Pitter KL, et al. Osteopontin-CD44 signaling in the glioma perivascular niche enhances cancer stem cell phenotypes and promotes aggressive tumor growth. Cell stem cell 2014;14:357-69 40. Maltzman JS, Carman JA, Monroe JG. Role of EGR1 in regulation of stimulus- dependent CD44 transcription in B lymphocytes. Mol Cell Biol 1996;16:2283-94 41. Smith SM, Lyu YL, Cai L. NF-kappaB affects proliferation and invasiveness of breast cancer cells by regulating CD44 expression. PLoS One 2014;9:e106966 42. Smith SM, Cai L. Cell specific CD44 expression in breast cancer requires the interaction of AP-1 and NFkappaB with a novel cis-element. PLoS One 2012;7:e50867 43. To K, Fotovati A, Reipas KM, Law JH, Hu K, Wang J, et al. Y-box binding protein-1 induces the expression of CD44 and CD49f leading to enhanced self- renewal, mammosphere growth, and drug resistance. Cancer Res 2010;70:2840- 51 44. Kumazoe M, Takai M, Bae J, Hiroi S, Huang Y, Takamatsu K, et al. FOXO3 is essential for CD44 expression in pancreatic cancer cells. Oncogene 2017;36:2643-54 45. Biamonti G, Catillo M, Pignataro D, Montecucco A, Ghigna C. The alternative splicing side of cancer. Semin Cell Dev Biol 2014;32:30-6 46. Gricius D, Kazlauskas A. Analysis of oncogene GLI1 protein expression levels between differing grades of astrocytoma. BIOLOGIJA 2014;60:142-7 47. Rimkus TK, Carpenter RL, Qasem S, Chan M, Lo HW. Targeting the Sonic Hedgehog Signaling Pathway: Review of Smoothened and GLI Inhibitors. Cancers 2016;8 48. Zhu H, Carpenter RL, Han W, Lo HW. The GLI1 splice variant TGLI1 promotes glioblastoma angiogenesis and growth. Cancer letters 2014;343:51-61 49. Godar S, Ince TA, Bell GW, Feldser D, Donaher JL, Bergh J, et al. Growth- inhibitory and tumor- suppressive functions of p53 depend on its repression of CD44 expression. Cell 2008;134:62-73

CHAPTER III

NEDD4-mediated degradation of TUSC2 promotes GBM development and

progression

Tadas K. Rimkus1, Dongqin Zhu1, Richard L. Carpenter1,#, Ivy Paw1, Austin Arrigo1,

Sherona Sirkisoon1, Daniel Doheny1, Noah Aguayo1, Guangxu Jin1,2, Boris Pasche1,2,3,

Waldemar Debinski1,2,3, Hui-Wen Lo1,2,3,*

1 Department of Cancer Biology, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

2 Comprehensive Cancer Center, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

3 Brain Tumor Center of Excellence, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA

# Current Address: Department of Biochemistry and Molecular Biology, Indiana

University School of Medicine-Bloomington, JH 308 1001 E. 3rd St., Bloomington, IN

47405

The manuscript was written through contributions of all authors. T.K.R., D.Z., R.L.C.,

I.P., A.A., S.S., D.D., N.A., and G.J. developed the methodology and carried out

75 experiments. G.J. provided laboratory support and assistance with data analysis. G.J.,

B.C.P., and W.D. provided administrative and material support and review of the manuscript. T.K.R. and H.L. performed data analysis and wrote the manuscript. T.K.R. and H.L. conceived the research. All authors will review the manuscript prior to submission for publication.

The following chapter is formatted for submission to Cancer Research. Stylistic variations are due to journal requirements.

ABSTRACT

The molecular pathways driving glioblastoma (GBM) initiation and progression are still poorly understood. We report here that tumor suppressor candidate 2 (TUSC2) mRNA transcript is equally expressed in normal brain and GBM tissues, but TUSC2 protein is consistently lost in GBM cells and patient tissues. TUSC2 is expressed in all proposed

GBM cells of origin in normal brain tissue. Ubiquitin E3 ligase neural precursor cell expressed developmentally downregulated protein 4 (NEDD4) binds to and polyubiquitinates TUSC2 at lysine 71, leading to its proteasomal degradation. NEDD4 is highly expressed in GBM clinical samples, such that TUSC2 and NEDD4 are significantly inveresely expressed in GBM tissues, and NEDD4 expression is a negative prognostic marker for glioma patient survival. Re-expression of TUSC2 in GBM cells and glioma stem cells (GSCs) induces apoptosis and inhibits sphere formation in vitro, and orthotopic xenograft initiation and growth in vivo. Conversely, in TUSC2-expressing

GBM cells, downregulation of TUSC2 by siRNA and CRISPR/Cas9 promotes sphere formation and orthotopic xenograft growth. RNA-seq analysis of TUSC2-knockout GBM cells identified multiple genes affected by TUSC2. BCL2L1 (Bcl-XL) expression increases and BCL2L11 (Bim) expression decreases in response to TUSC2 downregulation. A

TUSC2 gene signature generated from RNA-seq data correlates with poor overall survival in GBM patients. Collectively, these findings advance our understanding of GBM biology by establishing TUSC2 as a novel tumor suppressor and prognostic marker in

GBM, and delineates a novel mechanism by which GBM cells downregulate TUSC2 expression.

INTRODUCTION

Glioblastoma (GBM) is a grade IV astrocytoma, the most common primary brain cancer in adults, and the most intractable glioma, associated with a dismal prognosis of 14-15 months (1,2). A vast majority of GBM (~90%) develop de novo without clinical or histological evidence of a less malignant precursor lesion. Genetically engineered mouse models and large-scale genome sequencing by TCGA have shown a number of oncogenes (EGFR, PDGFRA, PIK3CA, K-Ras, H-Ras, hTERT, and Akt) and tumor suppressors (p53, Rb, PTEN, p16INK4A/p14ARF, and NF1) to play roles in the development of GBM (3,4).

TUSC2 (tumor suppressor candidate 2; also known as FUS1) is a known lung cancer tumor suppressor (5). Lerman and Minna identified several genes, including TUSC2, on deleted chromosome 3p21.3 as putative lung cancer tumor suppressor genes (6).

3p21.3 deletion was later found to be rare in lung cancer (1.1%; TCGA) and most other cancer types, except for mesothelioma (36%) (7) and renal clear cell carcinoma (12%;

TCGA). No evidence of methylation was found in the TUSC2 gene promoter region in lung cancer samples (8). In contrast, TUSC2 promoter region was reported to be partially methylated in head and neck squamous cell carcinomas and normal salivary rinses, but unmethylated in normal mucosa (9). TUSC2 somatic mutations have not been found in any cancer specimens according to TCGA, although infrequent mutations have been reported in lung cancer cell lines (8). Despite the lack of TUSC2 deletion and promoter methylation, TUSC2 mRNA is frequently reduced in lung cancer (~80%), which has been attributed to transcriptional and post-transcriptional mechanisms, such as regulation by microRNAs (6,10,11).

The exact functions of TUSC2 remain unclear. Published studies suggest TUSC2 may have a wide range of functions. TUSC2 has been shown to induce G1 cell cycle arrest and apoptosis (8,12,13), regulate calcium signaling (14), modulate tyrosine kinase and

Ser/Thr kinase activity (13,15-18), affect gene expression (7), and enhance tumor- suppressive miRNA expression (19). TUSC2 protein is primarily localized in the mitochondria but can also be detected in the cytoplasm (6,20) Studies indicate that N- myristoylation of the TUSC2 protein is required for its tumor suppressive role in lung cancer (21).

To date, there has been only one study of TUSC2 expression in gliomas reporting that

TUSC2 is expressed at higher levels in low-grade gliomas than high-grade gliomas; and that TUSC2 is linked to increased miR-197 expression (19). A comprehensive study of

TUSC2 as tumor suppressor in GBM remains to be done. In this study, we find that

TUSC2 protein expression, but not mRNA, is frequently lost in GBM cell lines and patient samples compared to normal brain samples; and this is driven by increased proteasome-mediated degradation of TUSC2 protein. Additionally, we have found that

TUSC2 protein is expressed in the proposed cells of origin for GBM tumors. We discovered that the ubiquitin E3 ligase neural precursor cell expressed developmentally downregulated protein 4 (NEDD4) drives TUSC2 polyubiquitination and degradation in

GBM. Functionally, we show that re-expression of TUSC2 increased apoptosis, and decreased GBM neurosphere formation and orthotopic xenograft growth. TUSC2 depletion promoted sphere formation and GBM growth in vitro and in vivo. Additionally, we found that TUSC2 modulates expression of key cellular apoptotic proteins, including

Bim and Bcl-XL, underlining its role as a tumor suppressor in GBM. Our findings provide evidence that the tumor suppressor TUSC2 protein expression is lost in GBM via

NEDD4-mediated polyubiquitination and degradation; and that TUSC2 protein loss promotes increased malignancy of GBM tumors.

MATERIALS AND METHODS

Cell lines and patient specimens

NHA, C8-S, D54MG, U373MG, LN18, LN229, T98G, and U87MG cells were obtained from ATCC and cultured according to their reccomendations. Luciferase-expressing

G48aLL2 cells were developed by Dr. Waldemar Debinski (22). BTCOE 4525, 4536,

4710, 4795, and 4810 cell lines were developed and validated from patient tumors at the

Brain Tumor Center of Excellence (23). UC1 cells were a kind gift from Dr. Russell O.

Pieper at University of California San Fransisco. Cells were cultured as previously described (24). Patient-derived GSCs were kind gifts from Drs. Erik Sulman and Krishna

Bhat at University of Texas MD Anderson Cancer Center (25). GSCs were passaged under neurosphere-forming conditions in serum-free DMEM/F12 growth medium supplemented with B27, FGF, and EGF in order to preserve stem-like properties. Normal brain tissue microarray (BNC17011) and glioma tissue microarray (GL2083) were purchased from US Biomax. Additional GBM patient samples were from the Wake

Forest Brain Tumor Center of Excellence (23). We have obtained written informed consent from the patients; the studies were approved by Wake Forest insitutional review board and were conducted in accordance with recognized ethical guidelines.

Reagents pcDNA3.1-DYK-TUSC2 plasmid was purchased from Genscript (Ohu29763). pCI-HA-

NEDD4 plasmid was a kind gift from Dr. Joan Massague (Addgene). Trilencer-27 siRNAs targeting TUSC2 were purchased from Origene (#SR307768). CRISPR/Cas9 constructs, TUSC2-targeting guide RNA-1 (gRNA-1) 5’-CCCGCGGCGCGTGAATACGA-

3’, and guide RNA-2 (gRNA-2) 5’-CCCTTCGTATTCACGCGCCG-3’ were purchased from GenScript. NEDD4 shRNAs were purchased from Dharmacon (Cat #

V3SVHS00_4980307, V3SVHS00_5281333). Cycloheximide (CHX) was purchased from Santa Cruz Biotechnology, Inc (#sc-3508). Caspase inhibitor Z-VAD(OH)-FMK was purchased from Selleckchem (#S8102) and proteasome inhibitor MG132 was purchased from Calbiochem (#474791-5MG). Click-iT Plus TUNEL assay was purchased from

ThermFisher Scientific (#C10618).

Cell-free ubiquitination assay

Recombinant TUSC2 (Novus Biologicals; #NBP2-22910) and recombinant NEDD4

(Sigma-Aldrich; #SRP0226) were purchased. Additional reagents were purchased from

Boston Biochem. To analyze NEDD4-mediated TUSC2 ubiquitination, recombinant

TUSC2 (10 μM), E1 (100 nM; #E-304), E2 (1 μM; #E2-620), ubiquitin (~100 μM; #U-

100H), and NEDD4 (1 μM) were incubated in the presence of ATP and reaction buffer

(Boston Biochem) at 37°C for 1 hr. The reaction was terminated with sample buffer, boiled, and subjected to SDS-PAGE and immunoblotting for TUSC2 (Abcam; ab70182).

PTEN was previously reported as a target for NEDD4-mediated ubiquitination, so recombinant PTEN ((Cayman Chemical; #10009746) was used as a positive control and blotted for with PTEN antibody (Millipore; #04-035; 1:1000). (26)

Generation of isogenic stable cell lines

The TUSC2-targeting guide RNA-2 (gRNA-2) 5’-CCCTTCGTATTCACGCGCCG-3’ was cloned into the lentiviral vector pRSGCEH-U6-gRNA-EF1-Cas9-2A-Hygro by Cellecta

(Mountain View, CA). Plasmids were sequenced to confirm insertion of TUSC2-targeting gRNA. Lentiviral packaging was performed using a 3rd generation lentiviral pPACK packaging system from System Biosciences. The pPACK packaging plasmids along with

TUSC2 gRNA/Cas9 lentiviral-expressing plasmid were co-transfected into 293TN cells for 48-96 hrs and cell culture medium was subsequently collected. Viral particles were

82 concentrated, titers measured, and used to infect G48LL2 cells. Infected cells were selected for with hygromycin treatment. TUSC2 protein expression was confirmed by western blotting. Non-targeting control gRNA/Cas9 lentiviral particles were purchased from Sigma-Aldrich and used to infect G48LL2 cells.

The TUSC2 open reading frame (ORF) was cloned into the lentiviral vector pLVX-

TRE3G (Clontech). Plasmids were sequenced to confirm insertion of TUSC2 ORF.

Lentiviral packing was performed using 3rd generation lentiviral Lenti-X packaging system from Clontech. The Lenti-X packaging plasmids along with doxycycline-inducible pLVX-TRE3G-TUSC2 ORF-expressing plasmid were co-transfected in 293TN cells for

48-96 hrs and cell culture medium was subsequently collcted. Viral particles were concentrated, titers measured, and used to infect GSC-28 cells carrying the pLVX-Tet3G doxycycline-sensitive lentiviral vector. Infected cells were selected for with puromycin treatment. TUSC2 protein induction in response to doxycycline treatment was confirmed by western blotting.

Gene Set Enrichment Analysis (GSEA)

GSEA was performed by generating the Gene Matrix file (.gmx) by using signatures generated from RNA-seq of TUSC2 CRISPR knockout cells. The Gene Cluster Text file

(.gct) was generated from the TCGA GBM dataset. The Categorical Class file (.cls) was generated from patients in the TCGA GBM dataset based on a validated NEDD4 gene signature (27). The number of permutations for GSEA was set to 1,000 and we used the

TCGA gene list as the CHIP platform.

Animal studies

Animal studies were carried out as previously described (24). Briefly, female nude mice

6-8 weeks of age were used. Actively growing G48LL2 cells stably-expressing either

83 control gRNA/Cas9 lentivirus or TUSC2 gRNA/Cas9 lentivirus were injected at a concentration of 1-5 x 105 cells in 5 μL PBS into the right frontal lobe. Actively growing

GSC-28 cells stably-expressing the dox-inducible TUSC2 lentivirus were injected at a concentration of 1-5x105 cells in 5 μL PBS into the right frontal lobe. Mice were anesthetized with a ketamine/xylazine mixture and a burr hole was placed along the coronal suture through a scalp incision according to and approved IACUC protocol. Mice were treated with doxycycline hyclate (Sigma-Aldrich) administered at 2 mg/mL supplemented with 5% sucrose in the drinking water, maintained in darkened bottles and refreshed twice weekly. For bioluminescent imaging, xenograft-bearing mice were injected intraperitoneally with d-luciferin (PerkinElmer) at 100 mg/kg body weight and then imaged weekly using PerkinElmer IVIS100 imager.

Statistical analyses

Data are presented as mean±SE. Student’s T-test, one-way ANOVA, Chi-squared analysis, Fisher’s Exact test, and survival analyses were performed using GraphPad

Prism 5 and Sigma Plot version 11.0.

RESULTS

TUSC2 protein expression is frequently lost in GBM.

Whether TUSC2 is lost in GBM compared to normal brain tissues has not been investigated. Herein, we examined a panel of patient-derived GBM cell lines, established

GBM cell lines, GBM patient samples, and normal brain astrocytes for TUSC2 mRNA and protein expression. TUSC2 mRNA expression is frequently decreased in lung cancer patients (6,10). Interestingly, TUSC2 mRNA levels were unaffected, yet TUSC2 protein was frequently decreased in GBM samples compared to normal astrocytes (Fig.

1A). GEO data analysis (GSE4290) indicated that TUSC2 mRNA expression did not differ between normal brain samples and GBM patient samples, nor across glioma grades (Fig. 1B,C). Additionally, we examined TUSC2 gene promoter methylation using publically available TCGA data and found very little methylation of the TUSC2 gene promoter in GBM patients (Fig. 1D). We further examined GBM specimens and normal brain tissues for TUSC2 expression, and found TUSC2 protein to be frequently decreased or lost in GBM samples (Fig. 1E). Representative IHC images are shown in

Fig. 1F. Interestingly, we found TUSC2 protein to be expressed in Nestin-positive neural stem cells, glial fibrillary acidic protein (GFAP)-positive astrocytes, and Olig2-positive oligodendrocytes, all predicted cells of origin for GBM (Fig. 1G) (28,29). Together these results demonstrate that TUSC2 protein is expressed in normal brain tissues, including in the hypothesized cells of GBM origin, but is frequently decreased or lost in GBM.

TUSC2 protein is preferentially lost in GBM via proteasome-mediated degradation.

We asked what cellular mechanism drives TUSC2 protein loss in GBM. To determine differences in endogenous TUSC2 stability, we treated UC1 astrocytes and TUSC2- positive G48a GBM cells with protein synthesis inhibitor cycloheximide (CHX) for 0-12

85 hrs in order to track protein half-life. In G48a GBM cells, the half-life of TUSC2 protein was approximately 6 hours while the half-life of TUSC2 was not reached in astrocyte cells, indicating that TUSC2 protein is preferentially broken down in GBM cells compared to astrocytes (Fig. 2A). Increased degradation of TUSC2 in GBM cells was abrogated by treatment with proteasome inhibitor MG132, suggesting that TUSC2 protein is degraded by the ubiquitin-proteasome system. Next, we determined whether TUSC2 is ubiquitinated in GBM cells prior to degradation. We overexpressed TUSC2 in TUSC2- low U251MG cells, immunoprecipitated TUSC2 from these cells, and subjected the lysates to immunoblotting to examine levels of TUSC2 ubiquitination. We found that

U251MG cells overexpressing TUSC2 showed increased levels of ubiquitinated TUSC2 compared to vector control cells (Fig. 2B). Additionally, we showed that treatment with

MG132 leads to increased accumulation of ubiquitinated TUSC2 in GBM cells (Fig. 2C).

These results indicate that TUSC2 protein stability is decreased in GBM compared to normal astrocytes; and that TUSC2 is ubiquitinated prior to its proteasome-mediated degradation.

NEDD4 is overexpressed in GBM and binds to TUSC2.

In the ubiquitin-proteasome system, covalent attachment of multiple ubiquitin molecules to the target protein precedes its proteasomal degradation. This step is mediated by three key enzymes: the E1 ubiquitin-activating enzyme, the E2 ubiquitin-conjugating enzyme, and the E3 ubiquitin ligase. The E3 ubiquitin ligase is responsible for substrate specificty in the ubiquitin-proteasome system; and a large body of evidence exists indicating that deregulation of E3 ligases plays a role in cancer development (30). To identify the E3 ligase meditating TUSC2 ubiquitination, we analyzed a GEO data set

(GSE4290), consisting of mRNA expression profiles of 23 normal brains and 81 GBM samples, for differential gene expression of a panel of known E3 ligases (Fig. 2D). The

E3 ligases with a fold change > 2.0 are expanded in Fig. 2E. We next analyzed the ability of DTL, NEDD4, MDM2, and UBE3C to physically interact with TUSC2. We overexpressed TUSC2 in U373MG GBM cells and using immunoprecipitation followed by SDS-PAGE, found that when TUSC2 was immunoprecipitated, NEDD4 was the only of the candidate E3 ligases to co-immunoprecipitate with TUSC2 (Fig. 2F). Conversely, when we immunoprecipitated NEDD4, we found that TUSC2 co-immunoprecipitated with

NEDD4 (Fig. 2G). Together, these data show that NEDD4 is upregulated in GBM and is able to bind to TUSC2.

NEDD4 mediates ubiquitination and degradation of TUSC2 in GBM.

We next asked whether NEDD4 affects expression of TUSC2 in GBM cells. NEDD4 overexpression in TUSC2-positive G48a and U87MG GBM cells resulted in decreased expression of TUSC2, indicating that NEDD4 can modulate TUSC2 expression (Fig.

3A). Recombinant TUSC2 was polyubiquitinated by recombinant NEDD4 in a cell-free ubiquitination assay (Fig. 3B). PTEN, used a positive control, had been previously identified as a target for NEDD4-mediated ubiquitination (26). TUSC2 protein contains 6 lysine residues which can be targeted for ubiquitination. Mass spectrometry analysis of the resulting ubiquitinated recombinant TUSC2 product identified peptide fragments containing ubiquitinated lysines at positions 71, 84, and 93 (Fig. 3C, Supplemental

Figure 1). We generated lysine-to-arginine (KR) point mutations at these three lysine residues on TUSC2 and transfected the wild-type and mutant TUSC2 constructs into

TUSC2-negative GBM cells. Wild-type and mutant TUSC2 were immunoprecipitated and subjected to a cell-free ubiquitination assay using NEDD4 as the E3 ligase. We identified lysine 71 as the target for NEDD4-mediated polyubiquitination (Fig. 3D). We found that unlike wild-type TUSC2, the TUSC2-K71R mutant protein expression was not affected by NEDD4 overexpression (Fig. 3E). We also found that in GBM cells, NEDD4

87 knockdown using NEDD4-targeting shRNAs resulted in increased TUSC2 expression in these cells (Fig. 3F). Additionally, we found using a protein half-life assay that the

TUSC2-K71R mutant was more stable than the wild-type TUSC2 in transfected GBM cells (Fig. 3G). Together, these data showed that in GBM, NEDD4 ubiquitinates TUSC2 at lysine 71, leading to its proteasomal degradation.

TUSC2 and NEDD4 are inveresely expressed in GBM and normal brain tissues.

We next analyzed protein expression of TUSC2 and NEDD4 between astrocytes and

GBM cells. We found that TUSC2 protein is expressed at lower levels in patient-derived glioma stem cells (GSCs) and established GBM cell lines lines compared to cultured astrocytes (Fig. 4A). Additionally, NEDD4 was expressed at higher levels in GSCs and

GBM cells compared to normal astrocytes. Using immunohistochemistry, we examined

GBM specimens and normal brain tissues for NEDD4 expression, and found NEDD4 protein expression to be significantly increased in GBM patient samples (Fig. 4B).

Representative IHC images are shown in Fig. 4C. Chi-squared analysis of normal brain tissues and GBM patient samples indicated that NEDD4 and TUSC2 are significantly inversely expressed (Fig. 4D). Pair-wise analysis of GBM tissue samples indicate that

NEDD4 and TUSC2 are significantly differentially expressed in GBM patient samples

(Fig. 4E). TUSC2 mRNA expression is evenly expressed between GBM and normal brain samples, and we have shown a significant inverse correlation between TUSC2 and

NEDD4 (Fig. 1B,4D); therefore, we examined the prognostic value of NEDD4 expression as a surrogate marker for TUSC2 expression. In GBM patients only, NEDD4 mRNA expression was not correlated with poor patient survival; however, across multiple combined glioma datasets, we found that high NEDD4 expression is predictive of poorer overall survival in patients with gliomas (Fig. 4F-H) (31). Together, the results

88 in this figure indicate that TUSC2 and NEDD4 are inveresely expressed in GBM and normal brain tissues, and that NEDD4 is predictive of poorer glioma patient survival.

Restoring TUSC2 expression inhibits GBM growth

Previous studies indicated that re-expression of TUSC2 induces cell cycle arrest and apoptosis in lung cancer cells (8,12,13). Here, we found that TUSC2 overexpression interfered with the neurosphere-forming capability of GSCs (Fig. 5A,B). Additionally, we found that re-expression of TUSC2 in U373MG GBM cell lines inhibited their ability to form colonies and grow as neurospheres (Fig. 5C). Using a doxycycline (dox)-inducible system, we found that induction of TUSC2 protein expression by dox treatment (Fig. 5D) decreased the ability of GSC-28 GSCs (GSC28-indTUSC2) to form neurospheres (Fig.

5E). Additionally, we found that re-expression of TUSC2 in GSC-28 neurospheres increased apoptosis, as indicated by TUNEL assay staining (Fig. 5F). We next investigated the in vivo effects of TUSC2 re-expression on tumor-initiating capacity by treating mice with dox two days before orthotopic implantation of GSC28-indTUSC2 cells, and continuing dox treatment until the end of the study or treating mice after xenografts have already developed (Fig. 5G). TUSC2-based gene therapy has previously been evaluated in pre-clinical and clinical studies for patients with lung cancer, with some patients showing clinical benefit (12,15,17,32-36). To mirror this gene therapy, dox induction of TUSC2 was started in mice after appearance of an established tumor (Fig. 5G). In the pre-treatment group, TUSC2 re-expression in GSC-28 cells resulted in significantly imparied xenograft growth compared to controls (Fig. 5H,

Supplemental Figure 2). For the post-induction group, tumor growth inhibition trended towards significance (p=0.052). The tumor engraftment rate in the pre-treatment group was signficantly decreased in response to TUSC2 re-expression (Fig. 5I). We also found that TUSC2 re-expression prolonged survival of the GSC28-indTUSC2 xenograft-

89 bearing mice (Fig. 5J). Furthermore, we found increased apoptosis in TUSC2 re- expressing xenografts as indicated by TUNEL assay staining (Fig. 5K). The results in

Figure 5 provide evidence that TUSC2 acts as a tumor suppressor in GBM by inducing apoptosis and interfering with the tumor-forming capabilities of GBM cells in vitro and in vivo.

Loss of TUSC2 expression promotes GBM aggressiveness in vitro and in vivo.

To determine the functional role that TUSC2 protein loss plays in GBM, we knocked down TUSC2 expression using siRNA and CRISPR/Cas9 guide RNAs (gRNAs) in G48a

GBM cells. Knockdown of TUSC2 expression by siRNA (Fig. 6A) and gRNA (Fig. 6B) enhanced neurosphere fomation of G48a GBM cells. We further used a TUSC2- targeting guide RNA (gRNA) and CRISPR/Cas9 lentivirus to stably knockout TUSC2 expression in G48a GBM cells expressing empty vector or the oncogenic transcription factor tGLI1 and found that TUSC2 knockout cells retained an increased ability to form neurospheres in vitro (Fig. 6C,D). We had previously reported on the role of truncated glioma-associated oncogene homolog 1 (tGLI1) in promoting the stem cell-like phenotype in GBM and glioma stem cells (24,37-39). TUSC2 knockout also promoted growth of G48a-tGLI1 GBM intracranial xenografts in nude mice (Fig. 6E). These xenografts were not only larger, but also more proliferative, as indicated by Ki67 staining

(Fig. 6F-H). Together, these data demonstrate that TUSC2 functions as a tumor suppressor in GBM and that the loss of TUSC2 expression promotes an increased stem- like phenotype in vitro, and increased tumor growth and proliferation in vivo.

TUSC2 suppresses GBM growth by modulating cellular apoptotic machinery.

Next, we performed global RNA-seq of TUSC2 CRISPR versus control CRISPR GBM cells grown as neurospheres (Fig. 7A). TUSC2 mRNA was found to be significantly

90 decreased in TUSC2 CRISPR cells (See Appendix). Upon TUSC2 depletion, many genes were significantly (p<0.05) differentially expressed, including several members of the Bcl-2 family of proteins. Of interest, we found that the pro-apoptotic BCL2L11 (aka

Bim) was downregulated in TUSC2 CRISPR cells and upregulated in dox-treated

GSC28-indTUSC2 cells (Fig. 7B,C,F). Previous studies have indicated that Bim expression is upregulated in TUSC2-transfected mesothelioma cells and downregulated in mesothelioma tissue, which lacks TUSC2 protein expression, compared to healthy lung peritoneum, suggesting a conserved mechanism of TUSC2 tumor-suppressive function (7). Additionally, we found that the anti-apoptotic BCL2L1 (aka Bcl-XL) was upregulated in TUSC2 CRISPR cells and downregulated in dox-treated GSC28- indTUSC2 cells (Fig. 7D-F). Increased BCL2L1 expression in GBM tumors is significantly correlated with early patient death, while BCL2L11 expression has no significant prognostic value (Fig. 7G,H).

Furthermore, we asked whether TUSC2 expression can be used as a prognostic indicator for GBM . To indicate TUSC2 expression status, we generated gene signatures using the expression levels of the genes we identified as significantly upregulated or downregulated in our TUSC2 CRISPR GBM cells compared to control CRISPR cells

(p<0.05). We then analyzed TCGA datasets and found that the gene signature comprised of genes upregulated in response to TUSC2 deletion (N=683) correlated with worse overall survival (Fig. 7I). We found no significant prognostic value using the gene signature comprised of genes downregulated in respone to TUSC2 deletion (N=557)

(Fig. 7J). Kaplan-Meier curves generated using gene signatures of genes differentially upregulated or downregulated (p<0.01) show similar results (Supplemental Figure

3A,B). Using gene set enrichment analysis, we found that patients expressing high levels of a NEDD4 gene signature are enriched for the TUSC2 CRISPR upregulated

91 gene signature (aka TUSC2KO Upreg Signature), further validating our findings of the interplay between TUSC2 and NEDD4 (Fig. 7K,L). Patient stratification based on

NEDD4 mRNA levels did not indicate enrichment for the gene signatures

(Supplemental Figure 3C,D). Since we saw enrichment of the TUSC2KO Upreg gene signature in patients with high NEDD4 signature expression, we analyzed the prognostic signficance of these gene signatures together. We found that patients with high NEDD4 and high TUSC2KO Upreg gene signature expression had much poorer overall survival

(Fig. 7M). Additionally, high expression of a combined NEDD4 and TUSC2KO Upreg gene signature was also predictive of poorer GBM patient survival (Fig. 7N). Together, the data indicate that TUSC2 exerts its tumor suppressive role in GBM by altering expression of Bcl family of proteins and inducing apoptosis, and that this mechanism of tumor suppression may be conserved across cancer types. Additionally, the data provide evidence that TUSC2 expression may be an independent prognostic factor for patients with GBM.

DISCUSSION

We made the following important novel observations in this study: a) TUSC2 protein expression is frequently lost in GBM cells and patient samples, while mRNA levels are not affected; b) TUSC2 is preferentially degraded in GBM cells compared to normal astrocytes, and this effect is driven by NEDD4-mediated ubiquitination of TUSC2; c)

NEDD4 and TUSC2 are inversely expressed in GBM patient samples; d) re-expression of TUSC2 induces apoptosis in GBM and GSCs, and inhibits GSCs renewal in vitro and orthotopic tumor growth in vivo; e) knockout of TUSC2 expression in TUSC2-positive

GBM cells increases their neurosphere-forming capability and promotes orthotopic tumor growth in vivo; f) TUSC2 exerts its tumor suppressive role by altering global expression of key apoptotic mediators, including Bim and Bcl-XL. By reporting these findings, our study advances the biological understanding of the novel tumor suppressor TUSC2, the role of NEDD4 in regulating TUSC2 protein expression, and a key mechanism of

TUSC2-mediated tumor suppression in GBM.

Our molecular analyses revealed that TUSC2 protein expression is decreased or lost in the majority of GBM cases compared to normal tissues, a phenomenon first seen in lung cancer, but not comprehensively analyzed in GBM (19,40). Previous studies indicated a myristoylation-deficient mutant TUSC2 found in lung cancer is preferentially degraded compared to wild-type, but the mechanism has not been studied in GBM (21). We have discovered, for the first time, that E3ubiquitin ligase NEDD4 mediates proteasomal degradation of TUSC2 in GBM by facilitating polyubiquitination of lysine 71. Post- translational polyubiquitination and degradation of tumor suppressor gene products, such as p53, Rb, and p27, has been described in detail (41). MDM2 was found to preferentially ubiquitinate and degrade Rb following CDK4/6-mediated phosphorylation of Rb (42). An important future task is to explore the cellular events that signal TUSC2

93 for NEDD4-mediated ubiquitination and degradation in order to further understand the wider role of TUSC2 in tumor cell biology.

It is also important to investigate the potential interactions between TUSC2 and protein kinase signaling pathways that are commonly disregulated in GBM, such as EGFR and

PI3K/AKT, in light of the observations that re-expression of TUSC2 sensitizes lung cancer cells to EGFR and AKT inhibitors (15,17,18,43). Alongside our study, others have also shown that NEDD4 is overexpressed in GBM cells and patient tissues (44,45).

Recently, NEDD4 was reported to modulate EGFR and PI3K/AKT signaling in multiple tumor types (46,47). Whether loss of TUSC2 protein expression plays a role in the hyperactivation of these protein kinase signaling pathway in GBM is unknown, but is warranted as a future area of research.

In addition to its role in modulating protein tyrosine kinases, TUSC2 has also been shown to exert its tumor-suppressive function by inducing mitochondrial apoptosis in lung cancer cells (12,32,33). Previous studies have shown that re-expression of TUSC2 in lung cancer induces MDM2 downregulation and accumulation of p53 and APAF1, both key factors in mitochondrial-associated apoptosis (12,32). Additionally, TUSC2 can directly interact with APAF1 and regulate its activity (13,32). Since our results showed that TUSC2 modulates expression of Bcl-XL and Bim, it is possible that TUSC2 may interact with and effect the function of other key mediators of the intrinsic apoptosis signaling pathway. Further studies are needed to fully elucidate the mechanism of

TUSC2-mediated induction of apoptosis and how TUSC2 may interact with other proteins in this complicated and dynamic pathway.

Our present study uncovered that loss of TUSC2 expression in GBM results in global gene transcription alterations. Genome wide analysis of TUSC2 transcriptional effects in

94 malignant mesothelioma cells identified numerous pro- and anti-tumorigenic genes affected by TUSC2 re-expression (7). We showed that the TUSC2 knockout gene signature (genes upregulated in response to TUSC2 knockout GBM cells) predicted poor overall survival in patients with GBM. Interestingly, loss of TUSC2 protein expression has also been shown to be a significant negative prognostic factor in lung cancer patients (40). Comprehensive validation of TUSC2 protein as a prognostic marker in GBM still remains to be done. Identifying shared genes affected by TUSC2 is an important step in furthering our understanding of TUSC2 functionality that may be conserved across cancer types.

Loss of TUSC2 expression is frequently seen in cancerous tissues compared to normal tissues and associated pre-cancerous lesions (48,49). A previous study shows that

TUSC2 protein expression decreases with increasing glioma grade; however, the expression of TUSC2 in normal brain tissue was not studied (19). We found that TUSC2 protein is highly expressed in normal brain tissue and frequently lost in GBM cells and tissue samples. Additionally, we found that TUSC2 is expressed in all three proposed cells of origin for GBM, based on Nestin, GFAP, and Olig2 staining (28,29). The vast majority of GBM cases present at grade IV without evidence of precancerous lesions; therefore, understanding the role that TUSC2 loss plays in the malignant transformation of normal brain cells is key to the development of therapies that may target the therapeutic window presented by the lack of TUSC2 expression.

TUSC2 expression has been shown to be decreased in cancer by multiple mechanisms, including genomic deletion, loss of heterozygosity, and microRNA-mediated regulation; however, the expression pattern of TUSC2 in normal brain tissue and GBM has been poorly studied. It is also unknown what tumor-suppressive role TUSC2 plays in GBM. To address these gaps in knowledge, we initiated the current study and found that in GBM,

TUSC2 protein is frequently degraded by the E3 ubiquitin ligase NEDD4. Mechanistic studies provided evidence that TUSC2 re-expression induced global transcriptional changes and apoptosis in GBM, and that loss of TUSC2 expression in TUSC2-positive

GBM cells promoted orthotopic xenograft growth and proliferation. Translational studies indicated that NEDD4 and TUSC2 are inveresely expressed in GBM and normal brain tissues, and that a TUSC2-knockout gene signature may be used as a negative prognostic indicator of GBM patient survival. Our results established TUSC2 as a novel tumor suppressor and prognostic factor in GBM, and NEDD4 as a novel mediator of

TUSC2 protein expression.

FIGURES

Figure 1. TUSC2 protein is expressed in normal brain cells, but lost in GBM cells and patient samples. A) TUSC2 mRNA is equally expressed in astrocyte and GBM cells (top panel), but TUSC2 protein expression is signficantly decreased in GBM cell lines and patient samples (lower panel). B,C) TUSC2 is equally expressed between normal brain and across glioma grades. TUSC2 mRNA expression was determined for patient samples in GSE4290 dataset across normal healthy brain and glioma samples. D) TUSC2 promoter is very rarely methylated in GBM patient samples. TUSC2 gene promoter methylation was determined using TCGA datasets and calculating percentage of patients with methylated gene promoter. E) TUSC2 is highly expressed in normal brain samples, but not in GBM patient tumors. A cohort of normal healthy brain tissues (N=80) and GBM patient samples (N=63) were subjected to IHC using a TUSC2-specific antibody. Immunostained sections were scored by a pathologist to derive H-scores. F) Representative IHC images. G) TUSC2 is expressed in all predicted cells of GBM origin. Normal healthy mouse brain tissue was subjected to IF using antibodies specific to TUSC2, Nestin, GFAP, and Olig2. Overlayed images with yellow co-staining indicate coexpression of TUSC2 and lineage-specific markers. Student’s t-test and ANOVA were used to calculate p-values.

Figure 2. TUSC2 protein is preferentially degraded in GBM, and is bound to the E3 ligase NEDD4 in GBM cells. A) TUSC2 protein is preferentially degraded in GBM cells compared to normal astrocytes. To determine protein half-life, astrocyte and GBM cells were treated with 10 μg/mL cycloheximide (CHX) and either with or without proteasome inhibitor MG132 (10 μM). Densitometric analysis was completed using ImageJ and

99 plotted in the lower panel. B) TUSC2 is polyubiquitinated in GBM cells. TUSC2 was overexpressed in TUSC2-low U251MG cells and subjected to immunoprecipitation with anti-TUSC2 antibody. Resulting immunoprecipitates were subjected to SDS-PAGE and immunoblotted for ubiquitin. C) MG132 treatment inhibits degradation of polyubiquitinated TUSC2. U251MG cells were treated either with or without MG132 (10 μM). Resulting lysates were subjected to SDS-PAGE and immunoblotted for TUSC2. D) Multiple ubiquitin E3 ligases are overexpressed in GBM compared to normal brain samples. A GEO dataset (GSE4290) was analyzed for differential gene expression of a panel of known ubiquitin E3 ligases. Student’s t-test was used to compute p-values. E) Candidate ubiquitin E3 ligases with ~2 fold increased expression in GBM compared to normal brain samples. F) NEDD4 binds to TUSC2 in GBM cells. TUSC2 was overexpressed in U251MG cells and subjected to immunoprecipitation with anti-TUSC2 antibody. Resulting immunoprecipitates were subjected to SDS-PAGE and immunoblotted for DTL, NEDD4, MDM2, UBE3C, and TUSC2. G) TUSC2 is bound to NEDD4 in GBM cells. TUSC2 was overexpressed in U251MG cells and subjected to immunoprecipitation with anti-NEDD4 antibody. Resulting immunoprecipitates were subjected to SDS-PAGE and immunoblotted for NEDD4 and TUSC2.

100

Figure 3. TUSC2 protein is polyubiquitinated and degraded by NEDD4. A) NEDD4 overexpression decreases TUSC2 protein expression. NEDD4 was overexpressed in two different GBM TUSC2-positive GBM cell lines, G48a and U87MG, and the resulting lysates were subjected to SDS-PAGE and immunoblotted for NEDD4 and TUSC2. B) NEDD4 can ubiquitinate TUSC2 in a cell-free assay. Recombinant TUSC2 was incubated with or without NEDD4 in the presence of E1, E2, ubiquitin, and ATP, and

101 then subjected to SDS-PAGE and immunoblotting with anti-TUSC2 antibody. PTEN was used as a positive control for NEDD4 ubiquitination activity. C) Lysine 71 on TUSC2 is ubiquitinated by NEDD4. Ubiquitinated and un-ubiquitinated TUSC2 products from the NEDD4 cell-free ubiquitination assay were isolated from the SDS-PAGE gel and subjected to mass spectrometry analysis. Peptides from the ubiquitinated products containing lysine 71, 84, and 93 showed enrichment for markers indicative of ubiquitination (+114). D) Mutation of TUSC2 lysine 71 reverses NEDD4-mediated polyubiquitination of TUSC2. G48 TUSC2 CRISPR cells were transfected with wild-type Flag-TUSC2, mutant K71R, K84R, or K93R Flag-TUSC2 and subjected to immunoprecipitation with anti-Flag antibody-conjugated beads. Resulting immunoprecipitates were subjected to the cell-free ubiquitination assay with recombinant NEDD4. E) TUSC2-K71R expression is not affected by NEDD4 overexpression. U87MG cells were transfected with wild-type TUSC2 or mutant K71R TUSC2, and either NEDD4 or empty vector. Resulting lysates were subjected to SDS-PAGE and immunoblotting for TUSC2 and NEDD4. F) NEDD4 knockdown increases TUSC2 expression in GBM cells. U87MG cells were transduced with two non-overlapping shRNAs targeting NEDD4. Resulting lysates were subjected to SDS-PAGE and immunoblotting for NEDD4 and TUSC2. G) TUSC2-K71R mutation increases protein stability compared to wild-type TUSC2. G48 TUSC2 CRISPR cells were transfected with NEDD4, and either wild-type TUSC2 or mutant K71R TUSC2. Cells were treated with 10 μg/mL cycloheximide (CHX) and either with or without proteasome inhibitor MG132 (10 μM). Densitometric analysis was completed using ImageJ and plotted in the lower panel.

102

Figure 4. Inverse expression of TUSC2 and NEDD4 in GBM and normal brain samples. A) TUSC2 and NEDD4 are inversely expressed between cultured GBM and GSC cells, and normal astrocytes. Indicated cell lysates were subjected to SDS-PAGE and immunoblotted for TUSC2 and NEDD4. B) NEDD4 expression is increased in GBM tissue samples compared to normal brain samples. A cohort of normal healthy brain tissues (N=80) and GBM patient samples (N=63) were subjected to IHC using a NEDD4- specific antibody. Immunostained sections were scored by a pathologist to derive H-

103 scores. Student’s t-test was used to compute p-values. C) Representative IHC images of NEDD4 and TUSC2-stained tissues. D,E) TUSC2 and NEDD4 are inversely expressed in tissue samples. In Panel D, two-by-two Chi-squared analysis of brain and GBM tissues indicates that TUSC2 and NEDD4 are inversely expressed. Panel E shows pair- wise analysis of TUSC2 and NEDD4 expression in GBM patient samples. F-H) High NEDD4 mRNA expression predicts poorer overall survival in patients with gliomas. TCGA GBM Only (F), TCGA Combined Glioma (G) and Rembrandt Combined Glioma (H) datasets were analyzed and Kaplan-Meier plots were generated based on median expression of NEDD4 mRNA. Mantel-Cox log-rank test was used to compute p-values.

104

Figure 5. TUSC2 acts as a tumor suppressor in GBM. A,B) TUSC2 re-expression in GSCs inhibits their sphere-forming capability. GSCs were transfected with either TUSC2 or empty vector. TUSC2 expression was confirmed by SDS-PAGE and is shown in Panel A. 1000 GSCs were seeded and counted 7 days later. Results are shown in Panel

105

B. C) TUSC2 re-expression inhibits colony and sphere formation in cultured GBM cells. U251MG cells transfected with either TUSC2 or empty vector were seeded for colony formation assay (250 cells/well) or neurosphere formation assay (1000 cells/well). TUSC2 expression was confirmed by SDS-PAGE. D) Generation of doxycycline (dox)- inducible lentiviral system to re-express TUSC2 in GSC28 cells. Two separate clones were generated and dox-induction (1 μg/mL) of TUSC2 expression was confirmed by SDS-PAGE. E) Dox-induced re-expression of TUSC2 inhibits sphere-forming capacity of GSCs. GSC28-indTUSC2 cells were treated with dox (1 μg/mL) to induce TUSC2 expression, seeded for the assay, and counted after 7 days. F) TUSC2 re-expression induces apoptosis in GSCs. GSC28-indTUSC2 spheres were treated with dox (1 μg/mL) for 72 hours, attached to microscope slides using cytospin centrifugation, and subjected to TUNEL and DAPI staining. Spheres were visualized using a confocal microscope. G) Schema for GSC28-indTUSC2 orthotopic implantation animal experiment. Isogenic luciferase-expressing GSC28-indTUSC2 cells were injected into the right frontal lobe of female nude mice (N=10-12 per group) and tumor growth was assessed weekly via bioluminescent imaging. Pre-induction group mice started receiving water supplemented with dox (2 mg/mL) and 5% sucrose 2 days prior to intracranial cell implantation. Post- induction group started receiving water supplemented with dox (2 mg/mL) and 5% sucrose 7 days after intracranial implantation of cells. H) Re-expression of TUSC2 in GSC28-indTUSC2 orthotopic xenografts inhibits tumor growth. Bioluminescent images were analyzed and the mean total bioluminescent flux was plotted. I) TUSC2 re- expression impairs tumor development rates in GSC xenografts. Representative bioluminescent images of actively growing tumors at Day 21 are shown. Differences in tumor detection rate were determined using 2-by-3 Fisher’s exact test. J) TUSC2 re- expression prolongs survival in mice bearing GSC28-indTUSC2 xenografts. Kaplan- Meier survival graph is shown. Mantel-Cox log-rank test was used to determine p- values. K) TUSC2-expressing xenografts undergo increased apoptosis. Representative xenografts (N=5 per group) from each group were analyzed for apoptosis by TUNEL staining. Student’s t-test was used to compute p-values. All experiments were repeated at least three times to derive means and standard deviations.

106

Figure 6. TUSC2 downregulation promotes GBM growth in vitro and in vivo. A,B) TUSC2 knockdown promotes neurosphere formation in GBM cells. TUSC2-positive G48a GBM cells transfected with three distinct TUSC2-targeting siRNAs (A) or two

107 unique TUSC2 CRISPR/Cas9 guide RNA (gRNA) constructs (B) were subjected to the neurosphere assay. SDS-PAGE (top panel) and neurosphere assay results (bottom panels) are shown. C) Isogenic G48a GBM cells carrying lentiviral control CRISPR/Cas9 gRNA or TUSC2-targeting gRNA and either lentiviral mock or tGLI1 vector were subjected to SDS-PAGE and immunoblotting for TUSC2 and tGLI1 expression. D) Stable TUSC2 knockout increases sphere-forming capacity of GBM cells. Isogenic G48a GBM cells were subjected to the neurosphere assay. E) TUSC2 knockout rendered GBM orthotopic xenografts more aggressive in growth. Isogenic luciferase-expressing G48a cells were injected into the right frontal lobe of female nude mice (N=9 per group) and tumor growth was assessed weekly via bioluminescent imaging. Representative bioluminescent images of actively growing tumors at Day 56 are shown. F-H) TUSC2 knockout xenografts proliferate faster than control xenografts. Representative xenografts (N=5 per group) were subjected to H&E staining and IHC with indicated antibodies. Immunostained sections were scored by a pathologist and H-scores were calculated. Panel G shows representative images. Student’s t-test was used to compute p-values. All experiments were completed at least three times to derive means and standard deviation

108

Figure 7. TUSC2 modulation alters expression of apoptotic proteins in GBM cells. A) Global gene expression changes in response to TUSC2 knockout. Heatmap showing per-sample expression values of differentially expressed genes in G48a GBM cells following TUSC2 knockout by CRISPR/Cas9 (N=4) compared to control CRISPR/Cas9

(N=4). (p<0.05). B-F) Bim and Bcl-XL are inversely expressed in response to TUSC2

109 modulation. qPCR (B-E) and western blot (F) validation of BCL2L11 (Bim) and BCL2L1

(Bcl-XL) expression in G48a TUSC2 knockout cells and dox-induced GSC28-indTUSC2 cells. qPCR data are normalized to GAPDH and western blot data are normalized to α- tubulin. G,H) High Bcl-XL, but not Bim, expression is predictive of early GBM patient death. Kaplan-Meier survival graphs were generated using median BCL2L1 (Bcl-XL) and BCL2L11 (Bim) mRNA expression with data from the TCGA GBM dataset. I,J) Decreased TUSC2 expression is a negative prognostic indicator of overall survival in GBM patients. Kaplan-Meier survival curves were drawn with data from the TCGA GBM dataset using high or low expression of gene signatures generated from TUSC2 CRISPR RNA-seq analysis. TUSC2KO_Upreg signature contains genes (N=683) significantly upregulated (p<0.05) in response to TUSC2 CRISPR knockout. TUSC2KO_Downreg signature contains genes (N=557) significantly downregulated (p<0.05) in response to TUSC2 CRISPR knockout. K,L) GBM patients with high NEDD4 signature expression are enriched with TUSC2 knockout gene signature. GSEA was performed using gene signatures that are representative of genes upregulated or downregulated in response to TUSC2 knockout. Patients were stratified based on high or low NEDD4 signature expression using the TCGA GBM dataset. M,N) NEDD4 and TUSC2 knockout gene signatures are predictive of poor survival in GBM patients. Kaplan-Meier survival curves were drawn with data from the TCGA GBM dataset using high or low expression of NEDD4 and TUSC2 knockout gene signatures. Student’s t- test, Mantel-Cox log-rank test, and Wilcoxon log-rank test was used to calculate p-value. All experiments were completed at least three times to derive means and standard deviation.

110

Supplemental Figure 1. Mass spectrometry analysis of TUSC2 Lysine 84 and 93 ubiquitination signals. Ubiquitinated and un-ubiquitinated TUSC2 products from the NEDD4 cell-free ubiquitination assay were isolated from the SDS-PAGE gel and subjected to mass spectrometry analysis. Peptides from the ubiquitinated products

111 containing lysine 84 (A) and 93 (B) showed enrichment for markers indicative of mono- ubiquitination (+114).

Supplemental Figure 2. TUSC2 re-expression inhibited orthotopic glioma stem cell xenograft growth. Coronal sections of mouse brains bearing intracranial GSC28- indTUSC2 control (A), pre-induction (B), and post-induction (C) xenografts were subjected to hematoxylin and eosin staining, and subsequently imaged.

112

Supplemental Figure 3. TCGA GBM data analysis using TUSC2 CRISPR and NEDD4 gene signatures. Kaplan-Meier survival curves were drawn with data from the TCGA GBM dataset using high or low expression of gene signatures generated from TUSC2 CRISPR RNA-seq analysis. A) TUSC2KO_Upreg signature contains genes (N=411) significantly upregulated (p<0.01) in response to TUSC2 CRISPR knockout. B) TUSC2KO_Downreg signature contains genes (N=376) significantly downregulated

113

(p<0.01) in response to TUSC2 CRISPR knockout. C,D) TUSC2 knockout genes signatures are not enriched in GBM patients with high NEDD4 mRNA expression. GSEA was performed using gene signatures that are representative of genes upregulated or downregulated in response to TUSC2 knockout. Patients were stratified based on high or low NEDD4 mRNA expression using the TCGA GBM dataset. E) High expression of NEDD4 gene signature is predictive of poorer overall survival of GBM patients. Kaplan- Meier survival curves were drawn with data from the TCGA GBM dataset using high or low expression of the NEDD4 gene signature. Wilcoxon log-rank test was used to determine p-values.

114

SUPPLEMENTARY MATERIALS AND METHODS

Immunoblotting and Immunohistochemistry (IHC)

Immunoblotting and immunohistochemistry were performed as we previously described (24,39). Antibodies for immunoblotting include TUSC2 (Abcam; ab70182), β-actin (Cell Signaling Technology/CST; 8H10D10), NEDD4 (CST; C5F5), MDM2 (Santa Cruz Biotechnology/SCBT; SMP14), DTL (SCBT; B-8), UBE3C (SCBT; S-14), PTEN (Millipore; 04-035), and GLI1 (CST; L42B10). Antibodies for IHC included TUSC2 (ProteinTech; #11538-1-AP; 1:100), NEDD4 (Cell Signaling Technologies; #C5F5; 1:200), and Ki67 (NeoMarkers; #RB-9043-R7; 1:300). Histologic scores (H-Scores) were computed from both % positivity (A%, A=1-100) and intensity (B=0-3) using the equation, H-Score=A x B.

Immunofluorescence

Immunofluorescence staining was performed as previously described (50). Briefly, 10 μm-thick normal mouse brain slices were immunostained with antibodies including TUSC2 (ProteinTech; #11538-1-AP; 1:100), Nestin (Invitrogen; #MA1-110; 1:50), GFAP (Abcam; #ab10062; 1:1000), Olig2 (Invitrogen; #MA5-15810; 1:250), Alexa Fluor 488- conjugated secondary antibody (Invitrogen; #A11029; 1:500), and Alexa Fluor 594- conjugated secondary antibody (Invitrogen; #A11037; 1:500), stained with DAPI (Vector Labs; #H-1500), and then subjected to confocal microscopy. Images were overlayed using ImageJ.

Immunoprecipitation

Immunoprecipitation experiments were completed as previously described (51). Briefly, cells were lysed with SDS-free RIPA buffer supplemented with protease/phostphatase inhibitors followed by sonication and collection of supernatant. Whole cell extracts were pre-cleared with rabbit IgG-conjugated protein-A agarose beads. Cleared lysates were incubated with anti-Flag (Sigma-Aldrich; FLAGIPT1), anti-TUSC2 (Atlas Antibodies; HPA030116; 1:100), or anti-NEDD4 (Cell Signaling Technologies; #C5F5; 1:50) antibody overnight. Protein G-sepharose beads (Invitrogen) were then added and incubated for 2 hrs with agitation. Protein G-sepharose pellets were collected and washed. Washed pellets were boiled and subjected to SDS-PAGE and immunoblotting for DTL (Santa Cruz; #B-8; 1:100), NEDD4 (Cell Signaling Technologies; #C5F5; 1:1000), MDM2 (Santa Cruz; #SMP14; 1:200), UBE3C (Santa Cruz; #S-14; 1:200), and TUSC2 (Abcam; #ab70182; 1:1000).

RNA extraction and quantitative PCR

Total RNA was isolated using Promega SV Total RNA Isolation kit and subjected to quantitative PCR as previously described (50). The following primers were used:

BCL2L1 (aka Bcl-XL): Forward 5’-TGACCACCTAGAGCCTTGGA-3’ Reverse 5’- AAGAGTGAGCCCAGCAGAAC-3’ BCL2L11 (aka Bim): Forward 5’- AAGTTCTGAGTGTGACCGAG-3’ Reverse 5’-GCTCTGTCTGTAGGGAGGTA-3’.

115

Neurosphere assay

Adherent cells were seeded in neurosphere medium composed of base Neurobasal medium (LifeTechnologies) with 2% B27 (Sigma), 10 ng/mL FGF (Sigma), and 100 ng/mL EGF (Sigma). Cells were seeded into 24-well low-attachment plates (Corning) in neurosphere medium and spheres were counted 7-10 days later.

Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay TUNEL assay was performed per manufacturer’s instructions (ThermoFisher Scientific; C10618). Briefly, 10 μm-thick GSC28-indTUSC2 xenograft tissue slices or GSC28- indTUSC2 neurospheres attached to slides by Cytospin were stained per manufacturer’s instructions. Slides were mounted with DAPI (SouthernBiotech; #0100-020) and then subjected to confocal microscopy. Images were overlayed using ImageJ.

116

ACKNOWLEDGEMENTS

We are grateful for the assistance of Dr. Jingyun Lee from the Wake Forest

Comprehensive Cancer Center Proteomics and Metabolomics Shared Resource and

Cancer Genomics Shared Resource. We would also like to thank Eric Routh for assistance with data analysis. We would like to thank Drs. Erik Sulman and Krishna Bhat at University of Texas MD Anderson Cancer Center for gifting us the GSC cell lines. We also acknowledge the financial support from the Department of Defense; W81XWH-17-

1-0044 (to HWL) and NIH; R01-NS087169 (to HWL) and P30-CAO12197 (to BCP).

117

REFERENCES

1. Dolecek TA, Propp JM, Stroup NE, Kruchko C. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005-2009. Neuro-oncology 2012;14 Suppl 5:v1-49 2. Omuro A, DeAngelis LM. Glioblastoma and other malignant gliomas: A clinical review. JAMA 2013;310:1842-50 3. Brennan CW, Verhaak RGW, McKenna A, Campos B, Noushmehr H, Salama SR, et al. The Somatic Genomic Landscape of Glioblastoma. Cell 2013;155:462- 77 4. The Cancer Genome Atlas Research N, McLendon R, Friedman A, Bigner D, Van Meir EG, Brat DJ, et al. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 2008;455:1061 5. Rimkus T, Sirkisoon S, Harrison A, Lo H-W. Tumor Suppressor Candidate 2 (TUSC2; FUS-1) and Human Cancers. Discovery medicine 2017;23:325-30 6. Lerman MI, Minna JD. The 630-kb lung cancer homozygous deletion region on human chromosome 3p21.3: identification and evaluation of the resident candidate tumor suppressor genes. The International Lung Cancer Chromosome 3p21.3 Tumor Suppressor Gene Consortium. Cancer Res 2000;60:6116-33 7. Ivanova AV, Ivanov SV, Prudkin L, Nonaka D, Liu Z, Tsao A, et al. Mechanisms of FUS1/TUSC2 deficiency in mesothelioma and its tumorigenic transcriptional effects. Molecular cancer 2009;8:91 8. Kondo M, Ji L, Kamibayashi C, Tomizawa Y, Randle D, Sekido Y, et al. Overexpression of candidate tumor suppressor gene FUS1 isolated from the 3p21.3 homozygous deletion region leads to G1 arrest and growth inhibition of lung cancer cells. Oncogene 2001;20:6258-62 9. Demokan S, Chuang AY, Chang X, Khan T, Smith IM, Pattani KM, et al. Identification of guanine nucleotide-binding protein gamma-7 as an epigenetically silenced gene in head and neck cancer by gene expression profiling. International journal of oncology 2013;42:1427-36 10. Du L, Schageman JJ, Subauste MC, Saber B, Hammond SM, Prudkin L, et al. miR-93, miR-98, and miR-197 regulate expression of tumor suppressor gene FUS1. Molecular cancer research : MCR 2009;7:1234-43 11. Lee DY, Deng Z, Wang CH, Yang BB. MicroRNA-378 promotes cell survival, tumor growth, and angiogenesis by targeting SuFu and Fus-1 expression. Proc Natl Acad Sci U S A 2007;104:20350-5 12. Deng WG, Kawashima H, Wu G, Jayachandran G, Xu K, Minna JD, et al. Synergistic tumor suppression by coexpression of FUS1 and p53 is associated with down-regulation of murine double minute-2 and activation of the apoptotic protease-activating factor 1-dependent apoptotic pathway in human non-small cell lung cancer cells. Cancer Res 2007;67:709-17 13. Ji L, Roth JA. Tumor suppressor FUS1 signaling pathway. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer 2008;3:327-30 14. Uzhachenko R, Ivanov SV, Yarbrough WG, Shanker A, Medzhitov R, Ivanova AV. Fus1/Tusc2 is a novel regulator of mitochondrial calcium handling, Ca2+- coupled mitochondrial processes, and Ca2+-dependent NFAT and NF-kappaB pathways in CD4+ T cells. Antioxidants & redox signaling 2014;20:1533-47 15. Dai B, Yan S, Lara-Guerra H, Kawashima H, Sakai R, Jayachandran G, et al. Exogenous Restoration of TUSC2 Expression Induces Responsiveness to

118

Erlotinib in Wildtype Epidermal Growth Factor Receptor (EGFR) Lung Cancer Cells through Context Specific Pathways Resulting in Enhanced Therapeutic Efficacy. PLoS One 2015;10:e0123967 16. Lin J, Sun T, Ji L, Deng W, Roth J, Minna J, et al. Oncogenic activation of c-Abl in non-small cell lung cancer cells lacking FUS1 expression: inhibition of c-Abl by the tumor suppressor gene product Fus1. Oncogene 2007;26:6989-96 17. Meng J, Majidi M, Fang B, Ji L, Bekele BN, Minna JD, et al. The tumor suppressor gene TUSC2 (FUS1) sensitizes NSCLC to the AKT inhibitor MK2206 in LKB1-dependent manner. PLoS One 2013;8:e77067 18. Xiaobo C, Majidi M, Feng M, Shao R, Wang J, Zhao Y, et al. TUSC2(FUS1)- erlotinib Induced Vulnerabilities in Epidermal Growth Factor Receptor(EGFR) Wildtype Non-small Cell Lung Cancer(NSCLC) Targeted by the Repurposed Drug Auranofin. Scientific reports 2016;6:35741 19. Xin J, Zhang XK, Xin DY, Li XF, Sun DK, Ma YY, et al. FUS1 acts as a tumor- suppressor gene by upregulating miR-197 in human glioblastoma. Oncology reports 2015;34:868-76 20. Ivanova AV, Ivanov SV, Pascal V, Lumsden JM, Ward JM, Morris N, et al. Autoimmunity, spontaneous tumourigenesis, and IL-15 insufficiency in mice with a targeted disruption of the tumour suppressor gene Fus1. The Journal of pathology 2007;211:591-601 21. Uno F, Sasaki J, Nishizaki M, Carboni G, Xu K, Atkinson EN, et al. Myristoylation of the fus1 protein is required for tumor suppression in human lung cancer cells. Cancer Res 2004;64:2969-76 22. Debinski W, Gibo DM. Fos-related antigen 1 modulates malignant features of glioma cells. Molecular cancer research : MCR 2005;3:237-49 23. Ferluga S, Tome CM, Herpai DM, D'Agostino R, Debinski W. Simultaneous targeting of Eph receptors in glioblastoma. Oncotarget 2016;7:59860-76 24. Rimkus TK, Carpenter RL, Sirkisoon SR, Zhu D, Pasche B, Chan MD, et al. Truncated glioma-associated oncogene homolog 1 (tGLI1) mediates mesenchymal glioblastoma via transcriptional activation of CD44. Cancer research 2018 25. Bhat KPL, Balasubramaniyan V, Vaillant B, Ezhilarasan R, Hummelink K, Hollingsworth F, et al. Mesenchymal differentiation mediated by NF-kappaB promotes radiation resistance in glioblastoma. Cancer cell 2013;24:331-46 26. Wang X, Trotman LC, Koppie T, Alimonti A, Chen Z, Gao Z, et al. NEDD4-1 is a proto-oncogenic ubiquitin ligase for PTEN. Cell 2007;128:129-39 27. Sun J, Keim CD, Wang J, Kazadi D, Oliver PM, Rabadan R, et al. E3-ubiquitin ligase Nedd4 determines the fate of AID-associated RNA polymerase II in B cells. Genes & development 2013;27:1821-33 28. Lee JH, Lee JE, Kahng JY, Kim SH, Park JS, Yoon SJ, et al. Human glioblastoma arises from subventricular zone cells with low-level driver mutations. Nature 2018;560:243-7 29. Zong H, Verhaak RGW, Canoll P. The cellular origin for malignant glioma and prospects for clinical advancements. Expert review of molecular diagnostics 2012;12:383-94 30. Nakayama KI, Nakayama K. Ubiquitin ligases: cell-cycle control and cancer. Nature Reviews Cancer 2006;6:369 31. Carro A, Wang Q, Bowman RL, Verhaak RGW, Squatrito M. GlioVis data portal for visualization and analysis of brain tumor expression datasets. Neuro-oncology 2016;19:139-41

119

32. Deng WG, Wu G, Ueda K, Xu K, Roth JA, Ji L. Enhancement of antitumor activity of cisplatin in human lung cancer cells by tumor suppressor FUS1. Cancer gene therapy 2008;15:29-39 33. Ito I, Ji L, Tanaka F, Saito Y, Gopalan B, Branch CD, et al. Liposomal vector mediated delivery of the 3p FUS1 gene demonstrates potent antitumor activity against human lung cancer in vivo. Cancer gene therapy 2004;11:733-9 34. Li L, Yu C, Ren J, Ye S, Ou W, Wang Y, et al. Synergistic effects of eukaryotic coexpression plasmid carrying LKB1 and FUS1 genes on lung cancer in vitro and in vivo. Journal of cancer research and clinical oncology 2014;140:895-907 35. Lu C, Stewart DJ, Lee JJ, Ji L, Ramesh R, Jayachandran G, et al. Phase I clinical trial of systemically administered TUSC2(FUS1)-nanoparticles mediating functional gene transfer in humans. PLoS One 2012;7:e34833 36. Ren J, Yu C, Wu S, Peng F, Jiang Q, Zhang X, et al. Cationic liposome mediated delivery of FUS1 and hIL-12 coexpression plasmid demonstrates enhanced activity against human lung cancer. Current cancer drug targets 2014;14:167-80 37. Carpenter RL, Paw I, Zhu H, Sirkisoon S, Xing F, Watabe K, et al. The gain-of- function GLI1 transcription factor TGLI1 enhances expression of VEGF-C and TEM7 to promote glioblastoma angiogenesis. Oncotarget 2015;6:22653-65 38. Lo HW, Zhu H, Cao X, Aldrich A, Ali-Osman F. A novel splice variant of GLI1 that promotes glioblastoma cell migration and invasion. Cancer research 2009;69:6790-8 39. Zhu H, Carpenter RL, Han W, Lo HW. The GLI1 splice variant TGLI1 promotes glioblastoma angiogenesis and growth. Cancer letters 2014;343:51-61 40. Prudkin L, Behrens C, Liu DD, Zhou X, Ozburn NC, Bekele BN, et al. Loss and reduction of FUS1 protein expression is a frequent phenomenon in the pathogenesis of lung cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 2008;14:41-7 41. Kitagawa K, Kotake Y, Kitagawa M. Ubiquitin-mediated control of oncogene and tumor suppressor gene products. Cancer Science 2009;100:1374-81 42. Uchida C, Miwa S, Kitagawa K, Hattori T, Isobe T, Otani S, et al. Enhanced Mdm2 activity inhibits pRB function via ubiquitin-dependent degradation. The EMBO journal 2005;24:160-9 43. Brennan Cameron W, Verhaak Roel GW, McKenna A, Campos B, Noushmehr H, Salama Sofie R, et al. The Somatic Genomic Landscape of Glioblastoma. Cell 2013;155:462-77 44. Zhang H, Nie W, Zhang X, Zhang G, Li Z, Wu H, et al. NEDD4-1 regulates migration and invasion of glioma cells through CNrasGEF ubiquitination in vitro. PloS one 2013;8:e82789 45. Wang L, Zhu B, Wang S, Wu Y, Zhan W, Xie S, et al. Regulation of glioma migration and invasion via modification of Rap2a activity by the ubiquitin ligase Nedd4-1. Oncology reports 2017;37:2565-74 46. Weng M, Luo ZL, Wu XL, Zeng WZ. The E3 ubiquitin ligase NEDD4 is translationally upregulated and facilitates pancreatic cancer. Oncotarget 2017;8:20288-96 47. Shao G, Wang R, Sun A, Wei J, Peng K, Dai Q, et al. The E3 ubiquitin ligase NEDD4 mediates cell migration signaling of EGFR in lung cancer cells. Molecular cancer 2018;17:24 48. Liang S, Zhang N, Deng Y, Chen L, Zhang Y, Zheng Z, et al. miR-663b promotes tumor cell proliferation, migration and invasion in nasopharyngeal carcinoma through targeting TUSC2. Experimental and therapeutic medicine 2017;14:1095- 103

120

49. Liu F, Gong R, He B, Chen F, Hu Z. TUSC2P suppresses the tumor function of esophageal squamous cell carcinoma by regulating TUSC2 expression and correlates with disease prognosis. BMC Cancer 2018;18:894 50. Sirkisoon SR, Carpenter RL, Rimkus T, Anderson A, Harrison A, Lange AM, et al. Interaction between STAT3 and GLI1/tGLI1 oncogenic transcription factors promotes the aggressiveness of triple-negative breast cancers and HER2- enriched breast cancer. Oncogene 2018;37:2502-14 51. Carpenter RL, Han W, Paw I, Lo H-W. HER2 Phosphorylates and Destabilizes Pro-Apoptotic PUMA, Leading to Antagonized Apoptosis in Cancer Cells. PloS one 2013;8:e78836

121

CHAPTER IIV

TUSC2-downregulation and tGLI1-overexpression converge to promote GBM

development and progression

Tadas K. Rimkus1 and Hui-Wen Lo1,2,3

1 Department of Cancer Biology, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

2 Comprehensive Cancer Center, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA;

3 Brain Tumor Center of Excellence, Wake Forest School of Medicine, Medical Center

Boulevard, Winston Salem, NC, 27157, USA

All authors participated in aspects of study conception and design. T.K.R. performed the experiments. T.K.R. wrote the manuscript. H.W.L. assisted with writing and conceptual guidance.

122

MATERIALS AND METHODS

Cell lines and reagents

Luciferase-expressing G48LL2 cells were developed by Dr. Waldemar Debinski (1). UC1 cells were a kind gift from Dr. Russell O. Pieper at University of California San

Fransisco. Trilencer-27 siRNAs targeting TUSC2 were purchased from Origene.

CRISPR/Cas9 constructs, TUSC2-targeting guide RNA- (gRNA) 5’-

CCCTTCGTATTCACGCGCCG-3’ was purchased from GenScript.

Generation of isogenic stable cell lines

The open reading frame for tGLI1 and GLI1 were cloned into the lentiviral expressing vector pCDH-CMV-MCS-EF1-RFP-Puro (CD516B-2) by System Biosciences (Palo Alto,

48-96 hrs and the cell culture medium was subsequently collected. Viral particles were concentrated, titers measured, and used to infect G48LL2 cells. Infected cells were treated with puromycin and FACS sorted for RFP-positive cells.

The TUSC2-targeting guide RNA-2 (gRNA-2) 5’-CCCTTCGTATTCACGCGCCG-3’ was cloned into the lentiviral vector pRSGCEH-U6-gRNA-EF1-Cas9-2A-Hygro by Cellecta

TUSC2 gRNA/Cas9 lentiviral-expressing plasmid were co-transfected into 293TN cells for 48-96 hrs and cell culture medium was subsequently collected. Viral particles were

123 concentrated, titers measured, and used to infect G48LL2 cells. Infected cells were selected for with hygromycin treatment. TUSC2 protein expression was confirmed by western blotting. Non-targeting control gRNA/Cas9 lentiviral particles were purchased from Sigma-Aldrich and used to infect G48LL2 cells.

Neurosphere assay

Cells were seeded in neurosphere medium composed of base Neurobasal medium

(LifeTechnologies) with 2% B27 (Sigma), 10 ng/mL FGF (Sigma), 100 ng/mL EGF

(Sigma), and 100 ng/mL sonic hedgehog (Sigma). Cells were seeded into 24-well low- attachment plates (Corning) in neurosphere medium and spheres were counted 7-14 days later.

Anchorage-independent colony formation assay

Cells were seeded in 6-well plates to determine anchorage-independent growth abilities.

Briefly, base agarose media (DMEM + 0.5% agarose) was heated to 40°C and 1.5 mL base agarose media was added to each well and allowed to cool for 30 minutes. Top agarose media was prepared (DMEM + 0.35% agarose) was heated to 40°C. 5000 cells were seeded per well in 1.5 mL of top agarose media. Fresh DMEM was added to wells every 3-4 days. Colonies were counted 10-14 days after seeding. tGLI1 activated signature (tGAS) and Gene Set Enrichment Analysis (GSEA) tGAS-7 was generated by averaging the median-centered expression of the seven tGLI1 target genes (CD24, VEGFA, VEGFC, HPA1, TEM7, VEGFR2, and CD44) (2-7). GSEA was performed by generating the Gene Matrix file (.gmx) by using signatures for genes upregulated or downregulated by TUSC2 CRISPR knockout (See Chapter III). The Gene

Cluster Text file (.gct) was generated from the TCGA GBM dataset. The Categorical

124

Class file (.cls) was generated based on the tGAS-7 score or a validated NEDD4 gene signature from patients in the TCGA GBM dataset (8). The number of permutations for

GSEA was set to 1,000 and we used the TCGA gene list as the chip platform.

Statistical analyses

Data are presented as mean±SE. The student’s t-test, one-way ANOVA, and Mantel-

Cox log-rank test were performed using Sigma Plot version 11.0 and GraphPad Prism 5.

125

RESULTS

Loss of TUSC2 expression is not sufficient to induce malignant transformation in astrocytes.

Whole-body TUSC2 knockout mice develop a unique immunological phenotype characterized by widespread inflammatory phenotypes and formation of spontaneous vascular and hematopoietic malignancies (9). Herein, we examined whether loss of

TUSC2 expression could induce malignant transformation in astrocytes. Studies have shown that gliomas can be generated by transforming astrocytes (10-14). We knocked down TUSC2 in immortalized astrocyte cells using siRNA and seeded cells in two different in vitro assays of malignant transformation. TUSC2 knockdown was confirmed by Western blot (Fig. 1A). Neither control-siRNA nor TUSC2-siRNA astrocyte cells formed spheres under serum-free neurosphere conditions (Fig. 1B). TUSC2 knockdown astrocytes formed more colonies in an anchorage-independent soft-agar colony assay; however, the results did not reach significance (p=0.15) (Fig. 1C). Together, this data shows that loss of TUSC2 expression alone is not sufficient to induce malignant transformation of immortalized astrocytes in vitro.

Multiple signaling pathways are enriched in GBM patients with low TUSC2 expression.

The Knudson’s two-hit hypothesis of cancer initiation states that two distinct alterations are required to cause a phenotypic change in normal cells (15). We hypothesized that activation of an oncogenic signaling pathway, in combination with TUSC2 downregulation, would cause malignant transformation in astrocytes. Using gene set enrichment analysis (GSEA) in an unbiased bioinformatic approach, we identified several oncogenic pathways that are enriched in patients expressing high levels of a

126 gene signature comprised of genes significantly upregulated in response to TUSC2 knockout (aka “TUSC2 low”) (Fig. 2A). We found that patients with high expression of the TUSC2-low gene signature were enriched for a gene signature indicative of tGLI1 activity (Fig. 2B). Interestingly, we found that a GLI1 activity signature was also enriched in these patients (Fig. 2A). We have previously shown that tGLI1 plays a key role in upregulating a highly malignant transcriptional program in GBM and breast cancer, resulting in increased tumor invasiveness, growth, angiogenesis, and stem-cell like phenotype (2-7). Additionally, we validated previous findings showing that TUSC2 plays a role in modulating EGFR and PI3K signaling (Fig. 2C,D) (16-18). Conversely, we found that the gene signature comprised of genes upregulated in response to TUSC2 knockout is enriched in patients with high expression of tGLI1 Activation Signature

(tGAS-7), which is made up of the validated genes shown to be regulated by gain-of- function tGLI1 (CD24, VEGFA, VEGFC, HPA1, TEM7, VEGFR2, CD44) (Fig. 2E). The data shown in Figure 2 suggest that tGLI1 overexpression may coordinate with TUSC2 downregulation to promote GBM development and progression. tGLI1 overexpression coordinates with TUSC2 downregulation to promote transformation of astrocytes in vitro. tGLI1 is a gain-of-function GLI1 transcription factor that, unlike the wild-type GLI1, is only expressed in cancerous tissue (2,5,6). To determine whether concomitant TUSC2 downregulation and tGLI1 overexpression could transform astrocyte, we transfected cells with empty vector, GLI1, or tGLI1, and then either with control or TUSC2-targeting siRNA. We confirmed GLI1 and tGLI1 overexpression, and TUSC2 knockdown by western blot (Fig. 3A). Transfected cells were then seeded in two different assays to determine their anchorage-independent growth capabilities. We found that only TUSC2- knockdown astrocytes overexpressing tGLI1 were capable of forming spheres (Fig. 3B).

127

Additionally, we found that these tGLI1-overexpressing and TUSC2-knockdown astrocytes formed the greatest number of colonies (ANOVA p=0.0007) in a soft-agar anchorage-independent growth assay (Fig. 3C). Together, these data indicate that tGLI1 overexpression and TUSC2 downregulation, which we have seen to be regular events in

GBM cells and tissues, but not normal brain tissue, promotes the malignant transformation of immortalized astrocytes in vitro. tGLI1 overexpression and TUSC2 downregulation promotes GBM progression.

We have previously shown that tGLI1 promotes a highly malignant phenotype in GBM by promoting a stem cell-like phenotype through transcriptional regulation of CD44 (6). We asked whether TUSC2 downregulation in TUSC2-positive GBM cells could further promote neurosphere formation in tGLI1-overexpressing cells. To address this, we transfected TUSC2-positive G48a GBM cells that stably express vector, GLI1, or tGLI1 with either a control or TUSC2-targeting siRNA. We confirmed knockdown of TUSC2 by western blot (Fig. 4A). We found that tGLI1-overexpressing TUSC2-knockdown cells formed the greatest number of spheres (ANOVA p=0.0063) (Fig. 4B, Supplemental

Fig. 1A). To further explore this phenomenon, we stably knocked out TUSC2 in G48a cells using a CRISPR/Cas9 system with a TUSC2-targeting guide RNA (gRNA) (Fig.

4C). Validating previous results, we found that tGLI1-overexpressing TUSC2-knockdown cells formed the greatest number of spheres (ANOVA p=0.0003) (Fig. 4D,

Supplemental Fig. 1B). Interestingly, in both the TUSC2 siRNA and CRISPR/Cas9 cells, we found that TUSC2-knockdown increased the sphere-forming capability of vector- and tGLI1-overexpressing cells, but not GLI1-overexpressing cells. These data indicate that TUSC2-downregulation and tGLI1-overexpression converge to promote

GBM progression in vitro.

128

DISCUSSION

There have been extensive studies into the genomic and mutational drivers of glioblastoma (GBM), yet the molecular mechanisms underlying GBM initiation remains poorly understood (19-22). Loss of TUSC2 expression is implicated in the pathogenesis of several solid and hematological cancers (9,23). In this study, we found that a) TUSC2 downregulation alone is not sufficient to transform immortalized astrocytes into glioma cells; b) several oncogenic signaling pathways, including tGLI1-mediated signaling, may coordinate with TUSC2 downregulation to drive GBM initiation; c) concomitant tGLI1 overexpression and TUSC2 knockdown promotes malignant transformation of astrocytes in vitro; d) tGLI1 overexpression in TUSC2 knockout GBM cells promotes an increased stem cell-like phenotype. By reporting these findings, our study advances the biological understanding of the interplay between novel oncogenes and tumor suppressors in

GBM, and sheds light on a potential novel mechanism of gliomagenesis.

In our study, TUSC2 downregulation alone did not transform immortalized astrocyte cells into glioma cells. Previous studies utilizing an immortalized astrocyte model demonstrated that introduction of oncogenes such as mutant H-ras and FOXM1B could transform these cells into gliomas; however, several well-characterized oncogenes such as AKT and EGFR failed to induce malignant transformation (11,13). An important future task would be to explore the role of TUSC2 downregulation in moderating gliomagenesis in an in vivo model. Prior to the discovery of adult neural stem/progenitor cells

(NSCs/NPCs), astrocytes were thought to be the cell of origin for gliomas. More recent studies have shown that the fundamental astrocytic marker GFAP is also expressed by adult NSCs, calling into question the proposed astrocytic origin of GBM (24). Thus, it is essential to examine the role of TUSC2 in adults NSCs and whether TUSC2 downregulation in this normal self-renewing cell population can lead to GBM formation.

129

Our results also indicate that TUSC2 knockdown may interact with several of key oncogenic pathways in established GBM. TUSC2 has been shown to play a role in modulating tyrosine and serine/threonine kinase signaling (16,25,26). Aberrant kinase signaling is one of the core pathways of GBM tumor maintenance, as defined by The

Cancer Genome Atlas (21). Whether there is cross-talk between TUSC2 and these known oncogenic signaling cascades in GBM is unknown, but is warranted as a future area of research.

We identified that tGLI1 signaling coordinates with TUSC2 downregulation to promote malignant transformation of astrocytes and sphere formation in established GBM cells.

We had previously shown that tGLI1 is enriched in the stem cell-like population of GBM cells and that tGLI1 directly upregulates transcription of CD44, promoting sphere formation (6). The coordination between tGLI1 and TUSC2 in promoting GBM initiation and progression remains to be elucidated. Previous studies show significant transcriptional alterations in response to TUSC2 re-expression, which may provide a basis for the cross-talk between tGLI1 and TUSC2 (27). Furthermore, it is essential to study the interaction between tGLI1 and TUSC2 in promoting gliomagenesis in animal models. In summary, my thesis research established TUSC2 as a novel tumor suppressor and prognostic factor in GBM, identified NEDD4 as a negative reegulator of

TUSC2 protein expression, defined tGLI1 as a novel mediator of mesenchymal GSCs through activaitng CD44 expression, and identified the functional cooperation between

TUSC2 loss and tGLI1 gain as a potential mechanism for GBM development and progression.

130

FIGURES

Figure 1. TUSC2 downregulation alone does not induce malignant transformation. A) Western blot showing successful knockdown of TUSC2 protein expression by siRNA. B) Immortalized astrocytes transfected with control or TUSC2-targeting siRNA did not form spheres under serum-free neurosphere growth conditions. C) TUSC2 knockdown did not signficantly enhance anchorage-independent growth in immortalized astrocytes. Cells transfected with control or TUSC2-targeting siRNA did not form more colonies in soft agar. Student’s t-test was used to calculate p-values. All experiments were completed at least three times to derive means and standard deviation.

131

Figure 2. A bioinformatics approach identifies multiple oncogenic pathways enriched in TUSC2 low GBM patients. A) The TCGA GBM dataset (N=165) was analyzed by gene set enrichment analysis for gene signature encompassing multiple oncogenic signaling pathways. Normalized enrichment score (NES) and nominal p-value are presented. B-D) Enrichment plots for tGLI1 gene signature (B), EGFR gene signature (C), and PI3K gene signature (D) stratified based on expression of the TUSC2 low gene signature. E) A gene signature comprised of genes upregulated in response to TUSC2 knockout (TUSC2KO_Upreg Signature aka TUSC2 low) is enriched in patients with high tGLI1 Activation Signature (tGAS-7).

132

Figure 3. TUSC2-downregulation and tGLI1-overexpression coordinate to transform astrocytes. A) GLI1 and tGLI1 were overexpressed and TUSC2 was knocked down by siRNA in immortalized astrocytes. Western blot indicates GLI1 and tGLI1 overexpression and TUSC2 knockdown. B) tGLI1-overexpressing TUSC2- knockdown astrocytes form spheres in vitro. Transfected cells were seeded under serum-free neurosphere-forming conditions and counted after 14 days. C) tGLI1- overexpressing TUSC2-knockdown increases anchorage-independent growth capabilities of transformed astrocytes. Transfected cells were seeded in a soft-agar assay and counted 10 days later. Student’s t-test and two-way ANOVA were used to determine p-values. * p<0.05, *** p<0.001, NS, not significant. All experiments were completed at least three times to derive means and standard deviation.

133

Figure 4. TUSC2-downregulation and tGLI1-overexpression promotes GBM sphere growth. A) Control or TUSC2-targeting siRNA was transfected into G48a GBM cells stably expressing vector, GLI1, or tGLI1. Western blot confirmed overexpression of GLI1 or tGLI1 and knockdown of TUSC2. B) TUSC2 downregulation increases sphere- forming capacity of GBM cells. Transfected cells were seeded under serum-free neurosphere-forming conditions and counted after 7 days. C) G48a GBM cells stably expressing vector, GLI1, or tGLI1 were transfected with control or TUSC2-targeting CRISPR/Cas9 guide RNA (gRNA) construct. Western blot confirmed stable knockout of TUSC2. D) TUSC2 knockout promotes sphere formation in GBM cells. Cells were seeded under serum-free neurosphere-forming conditions and counted 7 days later. Student’s t-test and two-way ANOVA were used to determine p-values. * p<0.05, ** p<0.01, NS, not significant. All experiments were completed at least three times to derive means and standard deviation.

134

Supplemental Figure 1. TUSC2-downregulation and tGLI1-overexpression promotes GBM sphere growth. A) Representative images of neurospheres from control or TUSC2-targeting siRNA-transfected G48a GBM cells. B) Representative images of neurospheres from stable control or TUSC2-targeting CRISPR/Cas9 guide RNA G48a GBM cells.

135

ACKNOWLEDGEMENTS

This work was supported by grant NIH R01-NS087169 (to HWL) and by start-up funds provided by the Wake Forest School of Medicine Department of Cancer Biology. We thank Richard L. Carpenter for assistance in technique development and data collection.

136

REFERENCES

1. Ferluga S, Tome CM, Herpai DM, D'Agostino R, Debinski W. Simultaneous targeting of Eph receptors in glioblastoma. Oncotarget 2016;7:59860-76 2. Cao X, Geradts J, Dewhirst MW, Lo HW. Upregulation of VEGF-A and CD24 gene expression by the tGLI1 transcription factor contributes to the aggressive behavior of breast cancer cells. Oncogene 2012;31:104-15 3. Carpenter RL, Paw I, Zhu H, Sirkisoon S, Xing F, Watabe K, et al. The gain-of- function GLI1 transcription factor TGLI1 enhances expression of VEGF-C and TEM7 to promote glioblastoma angiogenesis. Oncotarget 2015;6:22653-65 4. Han W, Carpenter RL, Lo H-W. TGLI1 Upregulates Expression of VEGFR2 and VEGF-A, Leading to a Robust VEGF-VEGFR2 Autocrine Loop and Cancer Cell Growth. Cancer Hallmarks 2013;1:28-37 5. Lo H-W, Zhu H, Cao X, Aldrich A, Ali-Osman F. A novel splice variant of GLI1 that promotes glioblastoma cell migration and invasion. Cancer research 2009;69:6790-8 6. Rimkus TK, Carpenter RL, Sirkisoon S, Zhu D, Pasche BC, Chan MD, et al. Truncated Glioma-Associated Oncogene Homolog 1 (tGLI1) Mediates Mesenchymal Glioblastoma via Transcriptional Activation of CD44. Cancer Res 2018;78:2589-600 7. Zhu H, Carpenter RL, Han W, Lo H-W. The GLI1 splice variant TGLI1 promotes glioblastoma angiogenesis and growth. Cancer letters 2014;343:51-61 8. Sun J, Keim CD, Wang J, Kazadi D, Oliver PM, Rabadan R, et al. E3-ubiquitin ligase Nedd4 determines the fate of AID-associated RNA polymerase II in B cells. Genes & development 2013;27:1821-33 9. Ivanova AV, Ivanov SV, Pascal V, Lumsden JM, Ward JM, Morris N, et al. Autoimmunity, spontaneous tumourigenesis, and IL-15 insufficiency in mice with a targeted disruption of the tumour suppressor gene Fus1. The Journal of pathology 2007;211:591-601 10. Bachoo RM, Maher EA, Ligon KL, Sharpless NE, Chan SS, You MJ, et al. Epidermal growth factor receptor and Ink4a/Arf: convergent mechanisms governing terminal differentiation and transformation along the neural stem cell to astrocyte axis. Cancer Cell 2002;1:269-77 11. Dai B, Pieper RO, Li D, Wei P, Liu M, Woo SY, et al. FoxM1B regulates NEDD4- 1 expression, leading to cellular transformation and full malignant phenotype in immortalized human astrocytes. Cancer research 2010;70:2951-61 12. Sonoda Y, Ozawa T, Aldape KD, Deen DF, Berger MS, Pieper RO. Akt Pathway Activation Converts Anaplastic Astrocytoma to Glioblastoma Multiforme in a Human Astrocyte Model of Glioma. Cancer Research 2001;61:6674-8 13. Sonoda Y, Ozawa T, Hirose Y, Aldape KD, McMahon M, Berger MS, et al. Formation of Intracranial Tumors by Genetically Modified Human Astrocytes Defines Four Pathways Critical in the Development of Human Anaplastic Astrocytoma. Cancer Research 2001;61:4956-60 14. Uhrbom L, Kastemar M, Johansson FK, Westermark B, Holland EC. Cell type- specific tumor suppression by Ink4a and Arf in Kras-induced mouse gliomagenesis. Cancer Res 2005;65:2065-9 15. Berger AH, Knudson AG, Pandolfi PP. A continuum model for tumour suppression. Nature 2011;476:163-9 16. Dai B, Yan S, Lara-Guerra H, Kawashima H, Sakai R, Jayachandran G, et al. Exogenous Restoration of TUSC2 Expression Induces Responsiveness to

137

Erlotinib in Wildtype Epidermal Growth Factor Receptor (EGFR) Lung Cancer Cells through Context Specific Pathways Resulting in Enhanced Therapeutic Efficacy. PLoS One 2015;10:e0123967 17. Hood MI, Uzhachenko R, Boyd K, Skaar EP, Ivanova AV. Loss of mitochondrial protein Fus1 augments host resistance to Acinetobacter baumannii infection. Infection and immunity 2013;81:4461-9 18. Xiaobo C, Majidi M, Feng M, Shao R, Wang J, Zhao Y, et al. TUSC2(FUS1)- erlotinib Induced Vulnerabilities in Epidermal Growth Factor Receptor(EGFR) Wildtype Non-small Cell Lung Cancer(NSCLC) Targeted by the Repurposed Drug Auranofin. Scientific reports 2016;6:35741 19. Alcantara Llaguno S, Chen J, Kwon CH, Jackson EL, Li Y, Burns DK, et al. Malignant astrocytomas originate from neural stem/progenitor cells in a somatic tumor suppressor mouse model. Cancer Cell 2009;15:45-56 20. Brennan CW, Verhaak RGW, McKenna A, Campos B, Noushmehr H, Salama SR, et al. The somatic genomic landscape of glioblastoma. Cell 2013;155:462-77 21. The Cancer Genome Atlas Research N, McLendon R, Friedman A, Bigner D, Van Meir EG, Brat DJ, et al. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 2008;455:1061 22. Verhaak RGW, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, et al. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer cell 2010;17:98-110 23. Prudkin L, Behrens C, Liu DD, Zhou X, Ozburn NC, Bekele BN, et al. Loss and reduction of FUS1 protein expression is a frequent phenomenon in the pathogenesis of lung cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 2008;14:41-7 24. Doetsch F, Caille I, Lim DA, Garcia-Verdugo JM, Alvarez-Buylla A. Subventricular zone astrocytes are neural stem cells in the adult mammalian brain. Cell 1999;97:703-16 25. Lin J, Sun T, Ji L, Deng W, Roth J, Minna J, et al. Oncogenic activation of c-Abl in non-small cell lung cancer cells lacking FUS1 expression: inhibition of c-Abl by the tumor suppressor gene product Fus1. Oncogene 2007;26:6989-96 26. Meng J, Majidi M, Fang B, Ji L, Bekele BN, Minna JD, et al. The tumor suppressor gene TUSC2 (FUS1) sensitizes NSCLC to the AKT inhibitor MK2206 in LKB1-dependent manner. PLoS One 2013;8:e77067 27. Ivanova AV, Ivanov SV, Prudkin L, Nonaka D, Liu Z, Tsao A, et al. Mechanisms of FUS1/TUSC2 deficiency in mesothelioma and its tumorigenic transcriptional effects. Molecular cancer 2009;8:91

138

CHAPTER V

GENERAL DISCUSSION & FUTURE DIRECTIONS

Tadas K. Rimkus

139

TUSC2 and tGLI1 as ELEMENTS OF TRANSLATIONAL CANCER RESEARCH

Despite advances in treatment and understanding of the underlying biology, GBM remains a dismal disease. Deeper understanding of the molecular aberrations that occur in normal brain cells that eventually progress into established GBM lesions is key to developing novel therapies. The work presented in this dissertation outlines the pre- clinical rational for developing TUSC2- and tGLI1-based diagnostic and therapeutic tools for use in GBM.

TUSC2-Based Therapies in GBM

TUSC2 is being explored as a therapeutic option in non-small cell lung cancer (1-5).

Phase I/II clinical trials are underway in stage III/IV metastatic lung cancer patients using intravenous nanoparticle-mediated delivery of TUSC2 gene therapy, both as a monotherapy and in combination with the EGFR inhibitor erlotinib. Based on Phase I studies (NCT00059605), five out of 31 patients treated with TUSC2 gene therapy alone achieved stable disease, and robust increases of TUSC2 mRNA and protein expression were detected in matched pre- and post-treatment patient biopsies (2). Based on our findings that TUSC2 protein expression is consistently lost in GBM samples (Chapter

III), a TUSC2-based gene therapy could be a promising approach for the development of new therapies for GBM. Viral-mediated gene delivery of tumor suppressor genes into

GBM tumors has shown great efficacy in decreasing tumor burden in pre-clinical and clinical studies, and TUSC2 could be added to the current armamentarium of tumor suppressor genes being studied for this purpose (6).

Advances in the understanding of immune control of cancer and the subsequent development of immunomodulatory drugs has placed these therapies at the forefront of cancer treatment; however, efficacy of anti-PD1 or anti-PDL1 immune checkpoint

140 therapies in GBM in the adjuvant setting has only been reported in isolated case studies.

TUSC2 has been shown to play a unique role in modulating both the innate and antitumor immune responses, particularly in the context of immune checkpoint blockade

(4,7-11). Recent studies show that re-expression of TUSC2 in TUSC2-deficient non- small cell lung cancer cells downregulates expression of PD-L1 and sensitizes lung cancer xenografts to anti-PD1 therapy (4,7). Additionally, two recently published studies found that neoadjuvant anti-PD1 therapy extended patient survival in resectable and recurrent GBM (12,13). Thus, the combination of TUSC2 gene therapy and immune checkpoint blockade presents an interesting avenue of clinical research for the treatment of GBM.

Pre-clinical studies have shown TUSC2 gene therapy synergizes with anti-AKT, anti-

EGFR, and anti-mTOR therapies to reduce lung cancer burden in vivo (1,3,4). These combination therapies may be of particular interest for use in treating GBM, as EGFR and PI3K/AKT pathway alterations occur in 57% and 89.6% of GBM patients, respectively (14). However, tyrosine kinase inhibitor monotherapies continue to show limited clinical efficacy in this disease (15,16). As discussed previously, further studies into the mechanism of TUSC2-mediated kinase inhibition are required to develop the clinical relevance of these combination therapies, as well as to understand potential resistance mechanisms that may arise.

In addition to exploring therapeutic options based upon re-expression of TUSC2 in tumors, there is potential to exploit the phenotypic alterations to cancer cells caused by

TUSC2 loss in a synthetic lethality-based approach. Early studies in TUSC2-knockout mice revealed that TUSC2-deficient cells exhibited higher basal levels of mitochondrial reactive oxygen species (ROS) (10,11,17). Maintaining redox balance is critical for cancer cell proliferation and survival; thus, a therapeutic opportunity exists to target this

141 vulnerability presented by loss of TUSC2 (18). A recent study found that cancer cells with increased mitochondrial ROS were dependent on increased production of glutathione, the major cellular ROS scavenger (19). By therapeutically targeting enzymes in the glutathione synthesis pathway, the researchers were able to induce potent synthetic lethality and cell killing in tumor cells. This therapeutic strategy has been delivered clinically through multiple approaches, including NOV-002, a novel glutathione disulfide mimetic that synergizes with chemotherapy to inhibit tumor growth, invasion, and metastasis (20,21). The effectiveness of this approach in multiple tumor types suggests that this strategy may be applicable to TUSC2-deficient GBM tumors; however, this approach must first be investigated pre-clinically.

Targeting tGLI1-Postive Cancers

Clinical efficacy of SMO-targeted therapies has been mixed while GLI-targeted treatments are still in preclinical testing (22). Clearly, there is a need to deepen the biological understanding of tumors with aberrant SMO and GLI activities. To this end, our lab identified a truncated, gain-of-function isoform of the GLI1 transcription factor, tGLI1, which is present only in cancerous cells and is undetectable in normal cells, making it an ideal drug target (23-25). We have shown that tGLI1 is a gain-of-function

GLI1 that has a higher propensity than GLI1 to induce aggressive cancer phenotypes in both breast cancer and GBM, including increased growth, invasion, migration, angiogenesis, and stem-cell like properties (23-29). These observations point to the need to pharmacologically target tGLI1.

Currently, there is no means to specifically inhibit tGLI1. However, several tGLI1 target genes can be targeted using already developed agents, including CD24, VEGF-

A, VEGF-C, VEGFR2, HPA1 and TEM7. Studies have attempted to target CD24 with

142 a monoclonal antibody (mAb) in several cancer models, including lung, ovarian, colorectal, and pancreatic cancer (30,31). These studies showed decreased proliferation, motility, and tumorigenicity of cancer cell lines in mouse models, and one study showed specifically that combination therapy of CD24 mAb and gemcitabine strongly potentiated its anti-cancer efficacy in a mouse lung cancer model (31). Several steps towards marketing a CD24 mAb as a cancer therapeutic have been made. A Phase I clinical trial (NCT02650895) studying the safety profile of a CD24 mAb in healthy adults was recently completed. CD24 remains a viable target for targeting the HH-tGLI1 pathway and further refinement of immunotherapies could prove to be beneficial in treating tumors with tGLI1 expression.

Antiangiogenic therapies have been established as a new focus for cancer drug development, as increased tumor vascularization is indicative of aggressive cancer and required for growth and metastasis (32). Heparanase, also referred to as HPSE or HPA1, is another tGLI1 target gene that has been extensively researched as an anti-angiogenic, anti-cancer drug target (33). Several classes of heparanase inhibitors have been studied, the most characterized of which include the heparan sulfate mimetics PI-88 and PG545. Researchers have shown in myriad of cancer models that

PI-88 and PG545 potently inhibit tumor angiogenesis by binding the active site of heparanase, and that they can be used in combination with other chemotherapeutics as an adjuvant therapy (34-38). PI-88 has been tested in multiple Phase I, II, and II trials spanning solid tumors and hematological malignancies, while PG545 is currently being tested in a Phase I clinical trial for advanced solid tumors (NCT02042781). PI-

88 was fast-tracked for FDA approval in 2007 for treatment of post-resection hepatocellular carcinoma. Roneparstat, a heparin mimetic, has also been shown to

143 have strong anti-cancer therapeutic effects and is currently in a Phase I trial for multiple myeloma (NCT01764880) (38).

TEM7, also termed PLXDC1, has been used as a prognostic marker for resectable gastric and colorectal cancers. As currently understood, TEM7 does not appear to be a viable therapeutic target, though it retains usefulness as a key prognostic marker for progression, invasion, and metastasis in several types of solid tumors (39,40). There are no clinical trials currently testing experimental therapeutics targeting TEM7.

VEGF-A and VEGFRs are the most widely characterized targets for antiangiogenic therapeutics among tGLI1 target genes, and the drugs targeting them have shown a varying range of effectiveness across tumor types (41). There are currently two FDA- approved therapies for targeting VEGF-A and VEGFR interactions. Bevacizumab, trade name Avastin, is a humanized monoclonal antibody that inhibits VEGF-A by direct binding (42). It is currently approved as either first-line monotherapy or in combination for metastatic colorectal cancer, non-small cell lung cancer, metastatic renal cell carcinoma, GBM, and a variety of other advanced solid tumors (43). In

2011, the FDA revoked its approval for the use of bevacizumab in treating breast cancer, citing a lack of advantage in survival rates, no improvement in quality of life, and significant side effects (44). Currently, over 500 active clinical trials are using bevacizumab, both as a monotherapy and in combination, to treat a variety of solid tumors. Ziv-aflibercept, trade name ZALTRAP, is the other FDA-approved immunotherapy for targeting VEGF-A, though its mechanism of action is slightly different from bevacizumab. ZALTRAP is a recombinant fusion protein that acts as a decoy VEGF receptor, binding to VEGF-A and preventing it from interacting with

VEGFR-1 and VEGFR-2 (45). It was approved by the FDA in 2012 for use in combination therapy with FOLFIRI chemotherapeutic regimen for treatment of

144 metastatic colorectal cancer that has progressed following oxaliplatin treatment (46).

There are currently a multitude of active clinical trials testing ZALTRAP in a variety of solid tumors.

Our laboratory has shown that in addition to increasing VEGF-A expression, tGLI1 also increases expression of VEGF-C. The library of compounds for targeting VEGF-

C is not as extensively developed as the one for VEGF-A, although there is one immunotherapy showing promising results in clinical trials. Circadian Technologies created a human monoclonal antibody named AGX-100 that is currently undergoing a first-in-human Phase I clinical trial both as a monotherapy and in combination with bevacizumab for the treatment of metastatic solid tumors (NCT01514123) (47).

We have shown that tGLI1 is also able to upregulate transcription of VEGFR2, creating a powerful autocrine loop with VEGF-A (27). VEGFR2 has long been a key target in the development of antiangiogenesis drugs as binding of VEGF-A to

VEGFR2 accounts for the majority of the pro-angiogenic signals as observed in mouse models (48). There are currently nine FDA-approved drugs for targeting

VEGFR2, eight of which show multi-kinase inhibitory activity.

In addition to targeting tGLI1 downstream target genes, inhibiting tGLI1 synthesis and targeting its transcription co-regulators could also inhibit tGLI1 leading to tumor cell kill. Unfortunately, the splicing events leading to tGLI1 synthesis are still unknown.

Also unknown is whether tGLI1 requires transcription co-factors for its gain-of-function transcriptional functions. Whether tGLI1 is subjected to regulation by non-canonical pathways has not been investigated. Ideally, tGLI1 activity can be blocked by inhibiting the non-classical pathways that activate tGLI1 activity. Filling these

145 knowledge gaps of tGLI1 will help with developing strategies to target tGLI1-driven tumors (22).

146

FUTURE DIRECTIONS

The works outlined in this dissertation suggest that tGLI1 and TUSC2 play a role in the processes of normal cell transformation and GBM progression. It remains to be studied whether the presence of tGLI1 or absence of TUSC2 are driver or passenger events in GBM initiation. Additional work must be done to elucidate the roles of tGLI1 and TUSC2 in other aspects of tumor cell biology.

Roles of TUSC2 and tGLI1 in Gliomagenesis

We have shown that downregulation of TUSC2, in combination with tGLI1 expression, can transform astrocyte cells in vitro (Chapter IV); however, as mentioned previously, there is emerging evidence that the adult neural stem/progenitor cell (NSC/NPC) population is in fact the cell of origin for GBM (49-52). Adult NSCs/NPCs are distinguished from other resident brain cells by a variety of molecular markers, including glial fibrillary acidic protein (GFAP) and nestin (53). These studies used genetically engineered mouse models (GEMMs) of GBM generated by overexpressing oncogenes and/or deleting tumor suppressors in a cell type-specific manner. Cell-type specificity was achieved using a GFAP or nestin promoter transgene construct to drive Cre recombinase expression in the NSC/NPC population. Cre recombinase expression drives DNA recombination of transgenes flanked by LoxP sites (54).

Common genes studied in the context of GBM formation include the tumor suppressors NF1, p53, and PTEN, and the oncogene EGFRvIII (49,50).

In order to address the roles that TUSC2 and tGLI1 play in GBM formation, further experiments should be completed using normal neural stem cells isolated from adult brains. As these cells are the cells of origin for GBM, TUSC2 levels must first be analyzed in these cells to determine whether TUSC2 loss occurs in the process of

147 gliomagenesis. Furthermore, the effects of TUSC2 downregulation and tGLI1 overexpression, both alone and in combination, must be studied using in vitro malignant transformation assays outlined previously (Chapter IV). It is also necessary to generate our own transgenic mice where the expression of TUSC2 and tGLI1, both alone and in combination, can be controlled in a cell-type specific fashion. These mouse models would provide definitive evidence of the function of TUSC2 and tGLI1 in gliomagenesis.

Greater Understanding of tGLI1 in GBM Tumor Biology

We have previously shown that tGLI1 plays a gain-of-function role in regulating the transcription of genes, including CD24, VEGF-A, VEGF-C, VEGFR2, HPSE, and

TEM7, that promote increased tumor invasion, growth, and angiogenesis in GBM and breast cancer (23,24,26-28). In this dissertation, we outlined the role that tGLI1 plays in promoting the cancer stem cell population in GBM by regulating transcription of

CD44 (Chapter II) (25). We have seen that tGLI1 controls a transcriptional program that encompasses many of the hallmarks of cancer described by Hanahan and

Weinberg (32). However, whether tGLI1 can drive transcription of novel genes related to the other recognized hallmarks, including cancer cell metabolism, resistance to apoptosis, and immune cell evasion remains unknown. Hedgehog signaling through wild-type GLI1 has been shown to potently regulate lipid metabolism in mouse livers, but Hedgehog signaling, mediated either by GLI1 or tGLI1, has not been studied in regulating metabolism of any substrate in cancer cells (55). Additionally, GLI1 has been shown to regulate transcription of the potent anti-apoptotic protein Bcl-2; however, additional anti-apoptotic genes regulating by tGLI1 remain unidentified (56).

Recently, a study showed that Hedgehog signaling can upregulate expression of PD-

L1, dampening the anti-tumor T-cell response, providing evidence that tGLI1 may also

148 play a role in tumor immune evasion (57). These novel tGLI1 gene targets can be identified through the use of high-throughput RNA sequencing (RNAseq) to identify genes differentially expressed in tGLI1- and GLI1-overexpressing cells.

Our lab has also shown that tGLI1 and the wild-type GLI1 transcription factors both physically interact with signal transducer and activator of transcription 3 (STAT3) to upregulate transcription of novel genes that are predictive of poor overall survival in breast cancer patients (29). The Hedgehog signaling pathway has been shown to crosstalk with multiple oncogenic signaling cascades to promote cancer progression, including RAS/RAF/MEK/ERK, PI3K/AKT/mTOR, EGFR, Wnt, and Notch signaling pathways (58,59). It remains to be seen whether tGLI1 interacts with the terminal effector transcription factors of these oncogenic signaling pathways to promote transcription of novel target genes. Potential binding partners of tGLI1 can be discovered using immunoprecipitation of tGLI1 followed by mass spectrometry, and subsequent validation. Identifying these interacting proteins would provide key rationale to develop novel targeted combination therapies as a treatment for cancer.

TUSC2 Protein Loss and GBM Cell Physiology

TUSC2 has been implicated in a wide array of cellular processes in cancerous and non-cancerous cells (60). Yet, the exact mechanism of how TUSC2 exerts these functions remains unknown. Multiple studies have shown that TUSC2 is a cytoplasmic protein which resides in the region of the mitochondria (11,61,62). Recent studies indicate that TUSC2 regulates mitochondrial calcium homeostasis (10,11,62,63).

Reprogramming of intracellular calcium levels is crucial for cancer cells due to their high proliferation rates (64,65). Calcium acts as a key intracellular secondary messenger in multiple signaling pathways to control a vast number of biological

149 processes that are key to tumorigenic growth, including proliferation and cell death

(66). Elevated cytosolic calcium levels have been shown to activate the ERK and AKT pathways, coinciding with the activation of a pro-tumorigenic transcriptional program

(66,67). These findings point at a potential mechanism for the differential gene expression profiles we and other have seen in response to TUSC2 knockout or

TUSC2 re-expression; however, this line of research warrants further study (Chapter

III) (68). Furthermore, as TUSC2 re-expression increases mitochondrial apoptosis in multiple tumor types, additional research must be done to elucidate this mechanism

(61,69-71). Calcium is a critical sensitizing signal in inducing mitochondrial apoptosis

(72). We have shown that in GBM cells, TUSC2 knockout increases anti-apoptotic

Bcl-XL expression and reduces pro-apoptotic Bim expression (Chapter III). Bcl-XL is known to play a role in regulating fluctuations in cytoplasmic calcium levels, thus mediating apoptotic resistance (73). Additional studies must be completed to identify other intracellular factors that modulate apoptosis resistance in GBM cells in response to loss of TUSC2 protein expression.

150

CONCLUSION

Gaining a greater understanding of the molecular mechanisms that drive GBM initiation, progression and recurrence sets the basis for the development of more effective therapies to treat this dismal disease. The work presented in this dissertation contributes to the overall understanding of how tGLI1 overexpression and TUSC2 downregulation, both as singular factors and in combination, promote GBM initiation and progression. In Chapter II, I show that tGLI1, but not wild-type GLI1, is able to promote glioma stem cell growth and function by driving transcription of CD44.

Additionally I show that tGLI1 expression and activity is preferentially enriched in the mesenchymal subtype of GBM, a highly recurrent and radioresistant subtype of the disease. In Chapter III, I demonstrate the role that TUSC2 plays in suppressing GBM initiation and growth. I also provide evidence of a novel NEDD4-mediated mechanism of TUSC2 protein loss in GBM. This TUSC2 protein loss leads to greater GBM proliferation, growth, and progression. In Chapter IV, I outline a novel mechanism of

GBM initiation and progression. For the first time, we see that tGLI1 overexpression and TUSC2 downregulation combine to induce malignant transformation in astrocytes and increase sphere formation in GBM cells. Moving forward, robust animal models are required to further elucidate the functional cooperation between tGLI1 overexpression and TUSC2 downregulation in vivo. Furthermore, clinical translation of our biological findings will require optimization of existing treatment modalities through rational combination therapies. The work presented in this dissertation describes the roles of novel oncogenes and tumor suppressors in the processes of GBM initiation and progression, and furthers our understanding of the tumor cell biology that underlies these cellular processes.

151

REFERENCES

1. Dai B, Yan S, Lara-Guerra H, Kawashima H, Sakai R, Jayachandran G, et al. Exogenous Restoration of TUSC2 Expression Induces Responsiveness to Erlotinib in Wildtype Epidermal Growth Factor Receptor (EGFR) Lung Cancer Cells through Context Specific Pathways Resulting in Enhanced Therapeutic Efficacy. PLoS One 2015;10:e0123967 2. Lu C, Stewart DJ, Lee JJ, Ji L, Ramesh R, Jayachandran G, et al. Phase I clinical trial of systemically administered TUSC2(FUS1)-nanoparticles mediating functional gene transfer in humans. PLoS One 2012;7:e34833 3. Meng J, Majidi M, Fang B, Ji L, Bekele BN, Minna JD, et al. The tumor suppressor gene TUSC2 (FUS1) sensitizes NSCLC to the AKT inhibitor MK2206 in LKB1-dependent manner. PLoS One 2013;8:e77067 4. Meraz IM, Majidi M, Cao X, Lin H, Li L, Wang J, et al. TUSC2 Immunogene Therapy Synergizes with Anti-PD-1 through Enhanced Proliferation and Infiltration of Natural Killer Cells in Syngeneic Kras-Mutant Mouse Lung Cancer Models. Cancer immunology research 2018;6:163-77 5. Xiaobo C, Majidi M, Feng M, Shao R, Wang J, Zhao Y, et al. TUSC2(FUS1)- erlotinib Induced Vulnerabilities in Epidermal Growth Factor Receptor(EGFR) Wildtype Non-small Cell Lung Cancer(NSCLC) Targeted by the Repurposed Drug Auranofin. Scientific reports 2016;6:35741 6. Cai X, Sughrue ME. Glioblastoma: new therapeutic strategies to address cellular and genomic complexity. Oncotarget 2017;9:9540-54 7. Cao X, Zhao Y, Wang J, Dai B, Gentile E, Lin J, et al. TUSC2 downregulates PD- L1 expression in non-small cell lung cancer (NSCLC). Oncotarget 2017;8:107621-9 8. Ivanova AV, Ivanov SV, Pascal V, Lumsden JM, Ward JM, Morris N, et al. Autoimmunity, spontaneous tumourigenesis, and IL-15 insufficiency in mice with a targeted disruption of the tumour suppressor gene Fus1. The Journal of pathology 2007;211:591-601 9. Ren J, Yu C, Wu S, Peng F, Jiang Q, Zhang X, et al. Cationic liposome mediated delivery of FUS1 and hIL-12 coexpression plasmid demonstrates enhanced activity against human lung cancer. Current cancer drug targets 2014;14:167-80 10. Uzhachenko R, Issaeva N, Boyd K, Ivanov SV, Carbone DP, Ivanova AV. Tumour suppressor Fus1 provides a molecular link between inflammatory response and mitochondrial homeostasis. The Journal of pathology 2012;227:456-69 11. Uzhachenko R, Ivanov SV, Yarbrough WG, Shanker A, Medzhitov R, Ivanova AV. Fus1/Tusc2 is a novel regulator of mitochondrial calcium handling, Ca2+- coupled mitochondrial processes, and Ca2+-dependent NFAT and NF-kappaB pathways in CD4+ T cells. Antioxidants & redox signaling 2014;20:1533-47 12. Cloughesy TF, Mochizuki AY, Orpilla JR, Hugo W, Lee AH, Davidson TB, et al. Neoadjuvant anti-PD-1 immunotherapy promotes a survival benefit with intratumoral and systemic immune responses in recurrent glioblastoma. Nat Med 2019;25:477-86 13. Schalper KA, Rodriguez-Ruiz ME, Diez-Valle R, López-Janeiro A, Porciuncula A, Idoate MA, et al. Neoadjuvant nivolumab modifies the tumor immune microenvironment in resectable glioblastoma. Nature Medicine 2019;25:470-6 14. Brennan CW, Verhaak RGW, McKenna A, Campos B, Noushmehr H, Salama SR, et al. The somatic genomic landscape of glioblastoma. Cell 2013;155:462-77

152

15. Westphal M, Maire CL, Lamszus K. EGFR as a Target for Glioblastoma Treatment: An Unfulfilled Promise. CNS drugs 2017;31:723-35 16. Zhao H-F, Wang J, Shao W, Wu C-P, Chen Z-P, To S-ST, et al. Recent advances in the use of PI3K inhibitors for glioblastoma multiforme: current preclinical and clinical development. Molecular cancer 2017;16:100- 17. Yazlovitskaya EM, Voziyan PA, Manavalan T, Yarbrough WG, Ivanova AV. Cellular oxidative stress response mediates radiosensitivity in Fus1-deficient mice. Cell death & disease 2015;6:e1652-e 18. Liou G-Y, Storz P. Reactive oxygen species in cancer. Free radical research 2010;44:479-96 19. Garama DJ, Harris TJ, White CL, Rossello FJ, Abdul-Hay M, Gough DJ, et al. A Synthetic Lethal Interaction between Glutathione Synthesis and Mitochondrial Reactive Oxygen Species Provides a Tumor-Specific Vulnerability Dependent on STAT3. Molecular and cellular biology 2015;35:3646-56 20. Gumireddy K, Li A, Cao L, Yan J, Liu L, Xu X, et al. NOV-002, A Glutathione Disulfide Mimetic, Suppresses Tumor Cell Invasion and Metastasis. Journal of carcinogenesis & mutagenesis 2013;2013:S7-002 21. Townsend DM, He L, Hutchens S, Garrett TE, Pazoles CJ, Tew KD. NOV-002, a glutathione disulfide mimetic, as a modulator of cellular redox balance. Cancer Res 2008;68:2870-7 22. Rimkus TK, Carpenter RL, Qasem S, Chan M, Lo HW. Targeting the Sonic Hedgehog Signaling Pathway: Review of Smoothened and GLI Inhibitors. Cancers 2016;8 23. Cao X, Geradts J, Dewhirst MW, Lo HW. Upregulation of VEGF-A and CD24 gene expression by the tGLI1 transcription factor contributes to the aggressive behavior of breast cancer cells. Oncogene 2012;31:104-15 24. Lo H-W, Zhu H, Cao X, Aldrich A, Ali-Osman F. A novel splice variant of GLI1 that promotes glioblastoma cell migration and invasion. Cancer research 2009;69:6790-8 25. Rimkus TK, Carpenter RL, Sirkisoon S, Zhu D, Pasche BC, Chan MD, et al. Truncated Glioma-Associated Oncogene Homolog 1 (tGLI1) Mediates Mesenchymal Glioblastoma via Transcriptional Activation of CD44. Cancer Res 2018;78:2589-600 26. Carpenter RL, Paw I, Zhu H, Sirkisoon S, Xing F, Watabe K, et al. The gain-of- function GLI1 transcription factor TGLI1 enhances expression of VEGF-C and TEM7 to promote glioblastoma angiogenesis. Oncotarget 2015;6:22653-65 27. Han W, Carpenter RL, Lo H-W. TGLI1 Upregulates Expression of VEGFR2 and VEGF-A, Leading to a Robust VEGF-VEGFR2 Autocrine Loop and Cancer Cell Growth. Cancer Hallmarks 2013;1:28-37 28. Zhu H, Carpenter RL, Han W, Lo H-W. The GLI1 splice variant TGLI1 promotes glioblastoma angiogenesis and growth. Cancer letters 2014;343:51-61 29. Sirkisoon SR, Carpenter RL, Rimkus T, Anderson A, Harrison A, Lange AM, et al. Interaction between STAT3 and GLI1/tGLI1 oncogenic transcription factors promotes the aggressiveness of triple-negative breast cancers and HER2- enriched breast cancer. Oncogene 2018;37:2502-14 30. Sagiv E, Starr A, Rozovski U, Khosravi R, Altevogt P, Wang T, et al. Targeting CD24 for treatment of colorectal and pancreatic cancer by monoclonal antibodies or small interfering RNA. Cancer Res 2008;68:2803-12 31. Salnikov AV, Bretz NP, Perne C, Hazin J, Keller S, Fogel M, et al. Antibody targeting of CD24 efficiently retards growth and influences cytokine milieu in experimental carcinomas. British Journal Of Cancer 2013;108:1449

153

32. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144:646-74 33. Pisano C, Vlodavsky I, Ilan N, Zunino F. The potential of heparanase as a therapeutic target in cancer. Biochemical pharmacology 2014;89:12-9 34. Ferro V, Dredge K, Liu L, Hammond E, Bytheway I, Li C, et al. PI-88 and novel heparan sulfate mimetics inhibit angiogenesis. Seminars in thrombosis and hemostasis 2007;33:557-68 35. Ferro V, Liu L, Johnstone KD, Wimmer N, Karoli T, Handley P, et al. Discovery of PG545: a highly potent and simultaneous inhibitor of angiogenesis, tumor growth, and metastasis. Journal of medicinal chemistry 2012;55:3804-13 36. Liu CJ, Chang J, Lee PH, Lin DY, Wu CC, Jeng LB, et al. Adjuvant heparanase inhibitor PI-88 therapy for hepatocellular carcinoma recurrence. World journal of gastroenterology 2014;20:11384-93 37. Ostapoff KT, Awasthi N, Cenik BK, Hinz S, Dredge K, Schwarz RE, et al. PG545, an angiogenesis and heparanase inhibitor, reduces primary tumor growth and metastasis in experimental pancreatic cancer. Molecular cancer therapeutics 2013;12:1190-201 38. Zhang L, Ngo JA, Wetzel MD, Marchetti D. Heparanase mediates a novel mechanism in lapatinib-resistant brain metastatic breast cancer. Neoplasia (New York, NY) 2015;17:101-13 39. Burgos-Ojeda D, McLean K, Bai S, Pulaski H, Gong Y, Silva I, et al. A novel model for evaluating therapies targeting human tumor vasculature and human cancer stem-like cells. Cancer Res 2013;73:3555-65 40. Zhang ZZ, Hua R, Zhang JF, Zhao WY, Zhao EH, Tu L, et al. TEM7 (PLXDC1), a key prognostic predictor for resectable gastric cancer, promotes cancer cell migration and invasion. American journal of cancer research 2015;5:772-81 41. Sitohy B, Nagy JA, Dvorak HF. Anti-VEGF/VEGFR therapy for cancer: reassessing the target. Cancer Res 2012;72:1909-14 42. Shih T, Lindley C. Bevacizumab: an angiogenesis inhibitor for the treatment of solid malignancies. Clinical therapeutics 2006;28:1779-802 43. Keating GM. Bevacizumab: a review of its use in advanced cancer. Drugs 2014;74:1891-925 44. Couzin-Frankel J, Ogale Y. FDA. Once on 'fast track,' avastin now derailed. Science (New York, NY) 2011;333:143-4 45. Patel A, Sun W. Ziv-aflibercept in metastatic colorectal cancer. Biologics : targets & therapy 2014;8:13-25 46. Van Cutsem E, Tabernero J, Lakomy R, Prenen H, Prausova J, Macarulla T, et al. Addition of aflibercept to fluorouracil, leucovorin, and irinotecan improves survival in a phase III randomized trial in patients with metastatic colorectal cancer previously treated with an oxaliplatin-based regimen. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2012;30:3499-506 47. Baldwin ME, Tester AM, Phelan D, Klupacs R. Abstract LB-284: The novel therapeutic monoclonal antibody VGX-100 neutralises VEGF-C and inhibits tumor growth and metastasis in subcutaneous and orthotopic models of human cancer. Cancer Research 2011;71:LB-284-LB- 48. Kieran MW, Kalluri R, Cho YJ. The VEGF pathway in cancer and disease: responses, resistance, and the path forward. Cold Spring Harbor perspectives in medicine 2012;2:a006593

154

49. Alcantara Llaguno S, Chen J, Kwon CH, Jackson EL, Li Y, Burns DK, et al. Malignant astrocytomas originate from neural stem/progenitor cells in a somatic tumor suppressor mouse model. Cancer Cell 2009;15:45-56 50. Lee JH, Lee JE, Kahng JY, Kim SH, Park JS, Yoon SJ, et al. Human glioblastoma arises from subventricular zone cells with low-level driver mutations. Nature 2018;560:243-7 51. Alcantara Llaguno SR, Xie X, Parada LF. Cell of Origin and Cancer Stem Cells in Tumor Suppressor Mouse Models of Glioblastoma. Cold Spring Harbor Symposia on Quantitative Biology 2016;81:31-6 52. Alcantara Llaguno S, Sun D, Pedraza AM, Vera E, Wang Z, Burns DK, et al. Cell- of-origin susceptibility to glioblastoma formation declines with neural lineage restriction. Nature neuroscience 2019;22:545-55 53. Zhang J, Jiao J. Molecular Biomarkers for Embryonic and Adult Neural Stem Cell and Neurogenesis. BioMed research international 2015;2015:727542- 54. Eroshenko N, Church GM. Mutants of Cre recombinase with improved accuracy. Nature Communications 2013;4:2509 55. Matz-Soja M, Rennert C, Schonefeld K, Aleithe S, Boettger J, Schmidt-Heck W, et al. Hedgehog signaling is a potent regulator of liver lipid metabolism and reveals a GLI-code associated with steatosis. eLife 2016;5 56. Bigelow RL, Chari NS, Unden AB, Spurgers KB, Lee S, Roop DR, et al. Transcriptional regulation of bcl-2 mediated by the sonic hedgehog signaling pathway through gli-1. The Journal of biological chemistry 2004;279:1197-205 57. Chakrabarti J, Holokai L, Syu L, Steele NG, Chang J, Wang J, et al. Hedgehog signaling induces PD-L1 expression and tumor cell proliferation in gastric cancer. Oncotarget 2018;9:37439-57 58. Brechbiel J, Miller-Moslin K, Adjei AA. Crosstalk between hedgehog and other signaling pathways as a basis for combination therapies in cancer. Cancer Treatment Reviews 2014;40:750-9 59. Takebe N, Miele L, Harris PJ, Jeong W, Bando H, Kahn M, et al. Targeting Notch, Hedgehog, and Wnt pathways in cancer stem cells: clinical update. Nature reviews Clinical oncology 2015;12:445-64 60. Rimkus T, Sirkisoon S, Harrison A, Lo HW. Tumor suppressor candidate 2 (TUSC2, FUS-1) and human cancers. Discovery medicine 2017;23:325-30 61. Uno F, Sasaki J, Nishizaki M, Carboni G, Xu K, Atkinson EN, et al. Myristoylation of the fus1 protein is required for tumor suppression in human lung cancer cells. Cancer Res 2004;64:2969-76 62. Uzhachenko R, Shanker A, Yarbrough WG, Ivanova AV. Mitochondria, calcium, and tumor suppressor Fus1: At the crossroad of cancer, inflammation, and autoimmunity. Oncotarget 2015;6:20754-72 63. Uzhachenko R, Boyd K, Olivares-Villagomez D, Zhu Y, Goodwin JS, Rana T, et al. Mitochondrial protein Fus1/Tusc2 in premature aging and age-related pathologies: critical roles of calcium and energy homeostasis. Aging 2017;9:627- 49 64. Parekh AB. Mitochondrial regulation of store-operated CRAC channels. Cell calcium 2008;44:6-13 65. Watson R, Parekh AB. Mitochondrial regulation of CRAC channel-driven cellular responses. Cell calcium 2012;52:52-6 66. Boland ML, Chourasia AH, Macleod KF. Mitochondrial dysfunction in cancer. Frontiers in oncology 2013;3:292 67. Feldman B, Fedida-Metula S, Nita J, Sekler I, Fishman D. Coupling of mitochondria to store-operated Ca(2+)-signaling sustains constitutive activation

155

of protein kinase B/Akt and augments survival of malignant melanoma cells. Cell calcium 2010;47:525-37 68. Ivanova AV, Ivanov SV, Prudkin L, Nonaka D, Liu Z, Tsao A, et al. Mechanisms of FUS1/TUSC2 deficiency in mesothelioma and its tumorigenic transcriptional effects. Molecular cancer 2009;8:91 69. Deng WG, Kawashima H, Wu G, Jayachandran G, Xu K, Minna JD, et al. Synergistic tumor suppression by coexpression of FUS1 and p53 is associated with down-regulation of murine double minute-2 and activation of the apoptotic protease-activating factor 1-dependent apoptotic pathway in human non-small cell lung cancer cells. Cancer Res 2007;67:709-17 70. Ito I, Ji L, Tanaka F, Saito Y, Gopalan B, Branch CD, et al. Liposomal vector mediated delivery of the 3p FUS1 gene demonstrates potent antitumor activity against human lung cancer in vivo. Cancer gene therapy 2004;11:733-9 71. Kondo M, Ji L, Kamibayashi C, Tomizawa Y, Randle D, Sekido Y, et al. Overexpression of candidate tumor suppressor gene FUS1 isolated from the 3p21.3 homozygous deletion region leads to G1 arrest and growth inhibition of lung cancer cells. Oncogene 2001;20:6258-62 72. Kroemer G, Reed JC. Mitochondrial control of cell death. Nat Med 2000;6:513-9 73. Li C, Wang X, Vais H, Thompson CB, Foskett JK, White C. Apoptosis regulation by Bcl-x(L) modulation of mammalian inositol 1,4,5-trisphosphate receptor channel isoform gating. Proc Natl Acad Sci U S A 2007;104:12565-70

156

APPENDIX

TUSC2 CRISPR RNA-seq list of significantly upregulated/downregulated genes

Ensemble ID Gene Symbol log2 Fold Adjusted P- Change value ENSG00000058085.13 LAMC2 1.56 3.02E-08 ENSG00000124785.7 NRN1 1.51 1.31E-13 ENSG00000178776.4 C5orf46 1.24 1.04E-06 ENSG00000196604.10 POTEF 1.22 6.35E-11 ENSG00000242256.3 RN7SL57P 1.18 0.005847074 ENSG00000242066.3 RN7SL214P 1.06 0.001362793 ENSG00000244194.3 RN7SL218P 0.99 0.002522903 ENSG00000126467.9 TSKS 0.84 0.01183127 ENSG00000243954.3 RN7SL743P 0.81 0.022547304 ENSG00000140961.11 OSGIN1 0.81 2.52E-17 ENSG00000240961.3 RN7SL415P 0.79 0.005296337 ENSG00000112782.14 CLIC5 0.79 0.000716851 ENSG00000198216.9 CACNA1E 0.79 1.07E-05 ENSG00000183454.12 GRIN2A 0.78 1.60E-18 ENSG00000242251.3 RN7SL20P 0.76 0.012563116 ENSG00000264275.1 RN7SL753P 0.75 0.009240769 ENSG00000126947.10 ARMCX1 0.74 4.66E-06 ENSG00000196542.7 SPTSSB 0.73 0.026377506 ENSG00000110328.5 GALNT18 0.72 0.015332872 ENSG00000171889.3 MIR31HG 0.71 0.036474353 ENSG00000233532.4 LINC00460 0.70 0.007713388 ENSG00000112796.8 ENPP5 0.70 2.23E-07 ENSG00000260757.1 RP11-120K18.2 0.70 0.009877127 ENSG00000167995.14 BEST1 0.68 3.36E-05 ENSG00000136235.14 GPNMB 0.67 3.36E-17 ENSG00000007174.16 DNAH9 0.67 0.000325529 ENSG00000147027.3 TMEM47 0.66 1.15E-26 ENSG00000227959.1 RP11-276H7.2 0.65 0.013310609 ENSG00000244671.3 RN7SL280P 0.64 0.027517347 ENSG00000276104.1 FP671120.5 0.63 0.00907264 ENSG00000152092.14 ASTN1 0.63 1.09E-17 ENSG00000147231.12 CXorf57 0.63 2.99E-12 ENSG00000154620.5 TMSB4Y 0.63 0.004112515 ENSG00000259932.1 CTD- 0.63 0.006050217

157

2651B20.7 ENSG00000080561.12 MID2 0.62 0.000120702 ENSG00000169436.15 COL22A1 0.62 7.27E-05 ENSG00000235027.1 AC068580.6 0.61 0.005002473 ENSG00000249378.4 LINC01060 0.61 0.033220964 ENSG00000198692.8 EIF1AY 0.60 2.33E-11 ENSG00000129159.6 KCNC1 0.60 5.77E-10 ENSG00000265559.2 RN7SL652P 0.60 0.030917019 ENSG00000106483.10 SFRP4 0.59 4.54E-05 ENSG00000133138.18 TBC1D8B 0.59 3.18E-09 ENSG00000101210.9 EEF1A2 0.58 3.29E-21 ENSG00000244514.3 RN7SL125P 0.58 0.044122417 ENSG00000100987.13 VSX1 0.58 0.000718889 ENSG00000171522.5 PTGER4 0.57 2.36E-09 ENSG00000227195.7 MIR663AHG 0.57 0.027936577 ENSG00000078596.9 ITM2A 0.57 1.37E-19 ENSG00000106034.16 CPED1 0.56 6.52E-07 ENSG00000143194.11 MAEL 0.56 0.039336579 ENSG00000269984.1 RP11-362K14.5 0.56 0.011397293 ENSG00000166446.13 CDYL2 0.55 8.14E-06 ENSG00000198576.3 ARC 0.55 0.004542566 ENSG00000088538.12 DOCK3 0.55 0.000952635 ENSG00000223947.1 AC016738.4 0.54 0.001320804 ENSG00000163975.10 MFI2 0.54 0.000544892 ENSG00000006283.16 CACNA1G 0.54 0.015354674 ENSG00000185567.6 AHNAK2 0.54 0.001967509 ENSG00000171346.12 KRT15 0.54 0.008721692 ENSG00000167779.6 IGFBP6 0.53 2.49E-05 ENSG00000223485.1 RP11-417E7.1 0.53 0.001501647 ENSG00000204291.9 COL15A1 0.53 0.013707249 ENSG00000136999.4 NOV 0.52 1.59E-13 ENSG00000234432.4 RP11- 0.52 0.043606645 1275H24.1 ENSG00000235750.8 KIAA0040 0.52 0.019674894 ENSG00000151789.8 ZNF385D 0.52 0.003618668 ENSG00000137573.12 SULF1 0.52 6.45E-10 ENSG00000279500.1 RP11-21K12.2 0.52 0.047256109 ENSG00000168243.9 GNG4 0.51 1.01E-07 ENSG00000205426.9 KRT81 0.51 2.47E-07 ENSG00000103811.14 CTSH 0.51 6.47E-08

158

ENSG00000104953.17 TLE6 0.50 0.000143614 ENSG00000215417.9 MIR17HG 0.50 0.015331032 ENSG00000165194.13 PCDH19 0.50 0.005445786 ENSG00000157502.11 MUM1L1 0.50 0.029323339 ENSG00000012817.14 KDM5D 0.49 4.02E-13 ENSG00000082397.14 EPB41L3 0.49 1.61E-08 ENSG00000131002.10 TXLNGY 0.48 5.25E-14 ENSG00000110002.14 VWA5A 0.48 0.000394066 ENSG00000159860.6 TCAF2P1 0.48 0.028137354 ENSG00000186007.8 LEMD1 0.48 4.07E-09 ENSG00000273812.1 WI2-87327B8.2 0.48 0.001630955 ENSG00000186340.13 THBS2 0.48 1.83E-17 ENSG00000164181.12 ELOVL7 0.48 0.011418095 ENSG00000203497.2 PDCD4-AS1 0.47 0.017767069 ENSG00000197769.5 MAP1LC3C 0.47 2.53E-07 ENSG00000106078.16 COBL 0.47 1.99E-05 ENSG00000137825.9 ITPKA 0.46 0.016558787 ENSG00000127561.13 SYNGR3 0.46 3.27E-07 ENSG00000202515.1 VTRNA1-3 0.46 0.020104729 ENSG00000135744.7 AGT 0.45 0.025656917 ENSG00000196834.10 POTEI 0.45 0.011130147 ENSG00000063180.7 CA11 0.45 1.11E-15 ENSG00000164620.7 RELL2 0.45 0.007075038 ENSG00000218996.1 RP1-99E18.2 0.45 7.80E-08 ENSG00000016391.9 CHDH 0.45 0.000345622 ENSG00000118257.15 NRP2 0.44 2.30E-19 ENSG00000129151.7 BBOX1 0.44 1.11E-15 ENSG00000176595.3 KBTBD11 0.44 0.024355375 ENSG00000214575.8 CPEB1 0.44 0.016856576 ENSG00000188219.13 POTEE 0.44 0.011711227 ENSG00000213244.3 HIST2H3DP1 0.44 0.002775364 ENSG00000130513.6 GDF15 0.44 8.74E-08 ENSG00000148123.13 RP11-35N6.1 0.44 0.044290621 ENSG00000277959.1 RP11-45A17.2 0.43 0.02381396 ENSG00000250221.1 KRT8P32 0.43 0.014073036 ENSG00000124942.12 AHNAK 0.43 9.01E-08 ENSG00000165521.14 EML5 0.43 0.001677018 ENSG00000184260.5 HIST2H2AC 0.43 0.000168275 ENSG00000110723.10 EXPH5 0.42 0.001974752 ENSG00000089199.8 CHGB 0.42 0.014229856

159

ENSG00000198513.10 ATL1 0.42 0.006424515 ENSG00000188483.7 IER5L 0.42 1.01E-11 ENSG00000166900.13 STX3 0.42 1.12E-09 ENSG00000221869.4 CEBPD 0.42 0.023591571 ENSG00000184545.9 DUSP8 0.41 0.001501647 ENSG00000102466.14 FGF14 0.41 0.006069347 ENSG00000164509.12 IL31RA 0.41 5.12E-10 ENSG00000207233.1 SNORA37 0.41 0.010089975 ENSG00000248780.1 RP11-632F7.1 0.41 7.43E-07 ENSG00000136542.7 GALNT5 0.41 0.000304001 ENSG00000140332.14 TLE3 0.41 3.36E-06 ENSG00000175874.8 CREG2 0.40 6.49E-06 ENSG00000164176.11 EDIL3 0.40 6.44E-21 ENSG00000147036.10 LANCL3 0.40 2.28E-05 ENSG00000182871.13 COL18A1 0.40 8.12E-10 ENSG00000210135.1 MT-TN 0.40 0.001757616 ENSG00000214530.6 STARD10 0.40 0.003184466 ENSG00000138772.11 ANXA3 0.40 0.004897741 ENSG00000067048.15 DDX3Y 0.40 3.32E-20 ENSG00000147394.17 ZNF185 0.40 0.003070275 ENSG00000146555.17 SDK1 0.40 1.34E-06 ENSG00000185338.4 SOCS1 0.39 0.040946583 ENSG00000178573.6 MAF 0.39 1.32E-06 ENSG00000122644.11 ARL4A 0.39 2.32E-12 ENSG00000164850.13 GPER1 0.39 0.000111102 ENSG00000174482.9 LINGO2 0.39 0.002191869 ENSG00000090006.16 LTBP4 0.39 9.68E-07 ENSG00000140945.14 CDH13 0.39 1.43E-17 ENSG00000274021.1 RP11-823E8.3 0.39 0.012563116 ENSG00000253190.3 AC084082.3 0.38 0.000693006 ENSG00000181649.5 PHLDA2 0.38 3.06E-05 ENSG00000274452.1 U2 0.38 0.012811231 ENSG00000080031.8 PTPRH 0.38 0.000590314 ENSG00000138678.9 AGPAT9 0.38 0.00219952 ENSG00000197375.11 SLC22A5 0.38 0.01636245 ENSG00000105048.15 TNNT1 0.38 7.07E-07 ENSG00000183878.14 UTY 0.37 3.49E-10 ENSG00000184489.10 PTP4A3 0.37 0.04929534 ENSG00000142910.14 TINAGL1 0.37 0.040683065 ENSG00000157214.12 STEAP2 0.37 0.038002354

160

ENSG00000069011.14 PITX1 0.37 0.019627623 ENSG00000078114.17 NEBL 0.37 0.001092744 ENSG00000237523.1 LINC00857 0.37 0.009871301 ENSG00000137501.15 SYTL2 0.37 2.14E-07 ENSG00000224184.4 AC096559.1 0.37 0.001727052 ENSG00000275219.1 RNU2-1 0.37 3.39E-05 ENSG00000278591.1 RNU2-1 0.37 3.39E-05 ENSG00000274062.1 RNU2-1 0.37 3.39E-05 ENSG00000274862.1 RNU2-1 0.37 3.39E-05 ENSG00000276596.1 RNU2-1 0.37 3.39E-05 ENSG00000277903.1 RNU2-1 0.37 3.39E-05 ENSG00000278774.1 U2 0.37 3.39E-05 ENSG00000162490.6 DRAXIN 0.36 1.01E-07 ENSG00000184363.8 PKP3 0.36 0.020616133 ENSG00000207340.1 RNVU1-10 0.36 0.038765873 ENSG00000064989.11 CALCRL 0.36 7.68E-07 ENSG00000077279.15 DCX 0.36 0.020662506 ENSG00000149256.13 TENM4 0.36 1.80E-05 ENSG00000170624.12 SGCD 0.36 0.018245631 ENSG00000164099.3 PRSS12 0.36 1.41E-07 ENSG00000074527.10 NTN4 0.36 8.85E-11 ENSG00000196511.12 TPK1 0.36 0.013888978 ENSG00000082293.11 COL19A1 0.35 2.05E-05 ENSG00000183598.3 HIST2H3D 0.35 0.000485466 ENSG00000165434.7 PGM2L1 0.35 2.54E-13 ENSG00000117600.11 LPPR4 0.35 1.02E-10 ENSG00000198400.10 NTRK1 0.35 0.022608998 ENSG00000114374.11 USP9Y 0.35 1.69E-12 ENSG00000274585.1 RNU2-1 0.35 1.90E-05 ENSG00000275616.1 RNU2-1 0.35 1.90E-05 ENSG00000002587.8 HS3ST1 0.35 0.000128781 ENSG00000184347.13 SLIT3 0.35 0.008129571 ENSG00000120885.18 CLU 0.35 3.36E-17 ENSG00000173482.15 PTPRM 0.35 1.45E-06 ENSG00000010278.10 CD9 0.35 1.24E-05 ENSG00000202538.1 RNU4-2 0.35 1.92E-05 ENSG00000215006.4 CHCHD2P2 0.35 0.010886753 ENSG00000184270.4 HIST2H2AB 0.35 2.61E-05 ENSG00000142227.9 EMP3 0.35 7.80E-08 ENSG00000146530.10 VWDE 0.35 0.000109856

161

ENSG00000203812.2 HIST2H2AA4 0.34 4.27E-05 ENSG00000161011.18 SQSTM1 0.34 4.13E-05 ENSG00000170365.8 SMAD1 0.34 1.54E-06 ENSG00000163412.11 EIF4E3 0.34 0.002769705 ENSG00000175048.15 ZDHHC14 0.34 0.000100781 ENSG00000101955.13 SRPX 0.34 1.35E-09 ENSG00000163618.16 CADPS 0.34 2.97E-07 ENSG00000179163.11 FUCA1 0.34 1.03E-05 ENSG00000170421.10 KRT8 0.34 6.11E-16 ENSG00000225091.3 SNORA71A 0.34 0.02022681 ENSG00000102575.9 ACP5 0.34 0.004836941 ENSG00000154589.5 LY96 0.34 0.001747617 ENSG00000188522.13 FAM83G 0.34 3.32E-05 ENSG00000170604.4 IRF2BP1 0.34 0.000120802 ENSG00000162645.11 GBP2 0.34 2.20E-07 ENSG00000099139.12 PCSK5 0.34 0.00037649 ENSG00000115884.9 SDC1 0.34 1.59E-06 ENSG00000273709.1 RNU2-1 0.34 5.28E-05 ENSG00000274432.1 RNU2-1 0.34 5.28E-05 ENSG00000278048.1 RNU2-1 0.34 5.28E-05 ENSG00000142494.12 SLC47A1 0.34 0.000268255 ENSG00000166289.5 PLEKHF1 0.34 0.043965156 ENSG00000065717.13 TLE2 0.33 0.000720456 ENSG00000152402.9 GUCY1A2 0.33 3.78E-10 ENSG00000113212.6 PCDHB7 0.33 0.020286626 ENSG00000164236.10 ANKRD33B 0.33 0.012781728 ENSG00000125864.10 BFSP1 0.33 0.00091933 ENSG00000146540.13 C7orf50 0.33 2.72E-10 ENSG00000116183.9 PAPPA2 0.33 0.037789698 ENSG00000200087.1 SNORA73B 0.33 4.39E-05 ENSG00000280071.2 CH507-9B2.3 0.32 0.039435238 ENSG00000254012.1 RP11-546B8.5 0.32 0.028997811 ENSG00000102271.12 KLHL4 0.32 8.20E-10 ENSG00000130522.5 JUND 0.32 1.71E-05 ENSG00000273002.1 RP11- 0.32 0.021901916 336K24.12 ENSG00000254682.1 RP11-660L16.2 0.32 0.027936577 ENSG00000164761.7 TNFRSF11B 0.32 2.02E-06 ENSG00000102007.9 PLP2 0.32 8.80E-05 ENSG00000165171.9 WBSCR27 0.32 0.023131616

162

ENSG00000112414.13 GPR126 0.32 0.003819038 ENSG00000143369.13 ECM1 0.32 0.009279192 ENSG00000158710.13 TAGLN2 0.31 1.41E-12 ENSG00000003147.16 ICA1 0.31 0.000135085 ENSG00000156453.12 PCDH1 0.31 0.001573185 ENSG00000200237.1 SNORA70 0.31 0.000472673 ENSG00000256235.1 SMIM3 0.31 2.35E-07 ENSG00000167972.12 ABCA3 0.31 0.000419181 ENSG00000168306.11 ACOX2 0.31 0.005528085 ENSG00000254285.3 KRT8P3 0.31 0.000335812 ENSG00000152137.5 HSPB8 0.31 0.00043287 ENSG00000197747.7 S100A10 0.31 2.82E-08 ENSG00000170381.11 SEMA3E 0.31 2.80E-05 ENSG00000121089.4 NACA3P 0.31 0.012566627 ENSG00000134853.10 PDGFRA 0.31 1.57E-07 ENSG00000145247.10 OCIAD2 0.31 1.05E-09 ENSG00000198517.8 MAFK 0.31 0.002244675 ENSG00000090013.8 BLVRB 0.31 2.20E-07 ENSG00000152503.8 TRIM36 0.31 1.43E-07 ENSG00000183741.10 CBX6 0.31 2.02E-12 ENSG00000112139.13 MDGA1 0.31 0.007126764 ENSG00000164086.9 DUSP7 0.30 0.00907264 ENSG00000184254.15 ALDH1A3 0.30 2.63E-09 ENSG00000113211.5 PCDHB6 0.30 0.048518421 ENSG00000163251.3 FZD5 0.30 1.18E-06 ENSG00000276368.1 HIST1H2AJ 0.30 0.009232054 ENSG00000073060.14 SCARB1 0.30 0.008599749 ENSG00000115641.17 FHL2 0.30 7.76E-08 ENSG00000277918.1 U1 0.30 0.000537541 ENSG00000259494.1 MRPL46 0.30 0.013310609 ENSG00000233864.6 TTTY15 0.30 0.000930546 ENSG00000155974.10 GRIP1 0.30 0.000598624 ENSG00000100097.10 LGALS1 0.29 1.51E-10 ENSG00000143507.16 DUSP10 0.29 8.17E-06 ENSG00000075618.16 FSCN1 0.29 1.76E-07 ENSG00000133131.13 MORC4 0.29 0.004761628 ENSG00000212402.1 SNORA74B 0.29 0.013310609 ENSG00000232533.1 AC093673.5 0.29 0.009657822 ENSG00000137834.13 SMAD6 0.29 0.017925999 ENSG00000162415.6 ZSWIM5 0.29 0.01726422

163

ENSG00000154655.13 L3MBTL4 0.29 0.031873457 ENSG00000173137.10 ADCK5 0.29 0.001738101 ENSG00000169093.13 ASMTL 0.29 0.000389157 ENSG00000100767.14 PAPLN 0.29 0.007045399 ENSG00000159784.16 FAM131B 0.29 1.01E-07 ENSG00000206737.1 RNVU1-18 0.29 0.001092702 ENSG00000273768.1 RNU1-1 0.29 0.001058531 ENSG00000210196.2 MT-TP 0.29 0.004705439 ENSG00000115414.17 FN1 0.29 0.000788659 ENSG00000170545.15 SMAGP 0.29 0.002379415 ENSG00000177706.8 FAM20C 0.29 2.60E-07 ENSG00000057294.12 PKP2 0.29 0.001878551 ENSG00000162909.16 CAPN2 0.29 3.52E-14 ENSG00000176463.12 SLCO3A1 0.29 0.002666855 ENSG00000173221.12 GLRX 0.29 0.000537541 ENSG00000155034.17 FBXL18 0.28 0.011379444 ENSG00000253982.1 CTD-2336O2.1 0.28 0.013273244 ENSG00000206588.1 RNU1-28P 0.28 0.001969305 ENSG00000125726.9 CD70 0.28 0.002121944 ENSG00000200795.1 RNU4-1 0.28 0.00616155 ENSG00000089041.15 P2RX7 0.28 0.005629947 ENSG00000142619.4 PADI3 0.28 3.31E-08 ENSG00000169242.10 EFNA1 0.28 0.001569901 ENSG00000129667.11 RHBDF2 0.28 0.042350987 ENSG00000235084.3 CHCHD2P6 0.28 0.000630447 ENSG00000152082.12 MZT2B 0.28 0.006593009 ENSG00000123240.15 OPTN 0.28 5.81E-05 ENSG00000116260.15 QSOX1 0.28 2.42E-08 ENSG00000107819.12 SFXN3 0.28 6.35E-07 ENSG00000196411.8 EPHB4 0.28 7.88E-07 ENSG00000081803.14 CADPS2 0.28 5.58E-05 ENSG00000105974.10 CAV1 0.27 0.000248951 ENSG00000206585.1 RNVU1-7 0.27 0.002787049 ENSG00000207005.1 RNU1-2 0.27 0.002764172 ENSG00000207389.1 RNU1-4 0.27 0.002764172 ENSG00000206633.1 SNORA80B 0.27 0.019215401 ENSG00000130830.13 MPP1 0.27 0.017216423 ENSG00000183098.9 GPC6 0.27 0.000161491 ENSG00000131981.14 LGALS3 0.27 8.16E-05 ENSG00000162631.17 NTNG1 0.27 0.000423683

164

ENSG00000129824.14 RPS4Y1 0.27 8.39E-12 ENSG00000206596.1 RNU1-27P 0.27 0.00634104 ENSG00000141401.10 IMPA2 0.27 0.000917905 ENSG00000165891.14 E2F7 0.27 0.00036381 ENSG00000086300.14 SNX10 0.27 0.005139607 ENSG00000205643.9 CDPF1 0.27 0.006750901 ENSG00000153714.5 LURAP1L 0.27 0.00075177 ENSG00000206652.1 RNU1-1 0.27 0.003478657 ENSG00000116132.10 PRRX1 0.27 7.31E-05 ENSG00000188191.13 PRKAR1B 0.27 0.011532681 ENSG00000171223.5 JUNB 0.27 0.007726245 ENSG00000090975.11 PITPNM2 0.27 0.01183127 ENSG00000126458.3 RRAS 0.27 0.000190766 ENSG00000196866.2 HIST1H2AD 0.27 0.001981997 ENSG00000177054.12 ZDHHC13 0.27 0.000436607 ENSG00000256167.1 ATF4P4 0.27 0.000418986 ENSG00000136052.8 SLC41A2 0.26 0.046831398 ENSG00000103145.9 HCFC1R1 0.26 2.29E-06 ENSG00000112852.6 PCDHB2 0.26 0.00362189 ENSG00000244509.3 APOBEC3C 0.26 0.000314845 ENSG00000164949.6 GEM 0.26 9.68E-05 ENSG00000278463.1 HIST1H2AB 0.26 0.027037514 ENSG00000242802.5 AP5Z1 0.26 0.007319283 ENSG00000274735.1 FP236383.3 0.26 0.049236336 ENSG00000162733.15 DDR2 0.26 0.008391439 ENSG00000150782.10 IL18 0.26 0.000247213 ENSG00000110811.18 P3H3 0.26 2.48E-05 ENSG00000114480.11 GBE1 0.26 0.00087944 ENSG00000167543.14 TP53I13 0.26 0.000543099 ENSG00000120756.11 PLS1 0.26 0.004112515 ENSG00000278588.1 HIST1H2BI 0.26 0.003781968 ENSG00000230364.1 RPL4P3 0.26 0.000110856 ENSG00000143429.8 AC027612.6 0.26 0.023199805 ENSG00000175745.10 NR2F1 0.26 2.36E-07 ENSG00000163898.8 LIPH 0.26 0.010574244 ENSG00000123700.4 KCNJ2 0.26 1.00E-08 ENSG00000110628.12 SLC22A18 0.26 0.005585539 ENSG00000104368.16 PLAT 0.26 1.43E-05 ENSG00000100994.10 PYGB 0.25 5.40E-08 ENSG00000170006.10 TMEM154 0.25 1.65E-08

165

ENSG00000211448.10 DIO2 0.25 0.014895068 ENSG00000182667.13 NTM 0.25 5.59E-06 ENSG00000103241.6 FOXF1 0.25 0.004549809 ENSG00000197409.7 HIST1H3D 0.25 0.008625152 ENSG00000208892.1 SNORA49 0.25 0.010988951 ENSG00000274997.1 HIST1H2AH 0.25 0.04238796 ENSG00000108984.12 MAP2K6 0.25 1.22E-06 ENSG00000178202.11 KDELC2 0.25 1.94E-06 ENSG00000260260.1 SNHG19 0.25 0.001789542 ENSG00000137968.15 SLC44A5 0.25 3.93E-08 ENSG00000100300.16 TSPO 0.25 0.000386307 ENSG00000145147.18 SLIT2 0.25 0.014846651 ENSG00000141682.11 PMAIP1 0.25 5.28E-05 ENSG00000141526.13 SLC16A3 0.25 0.017925999 ENSG00000057704.9 TMCC3 0.25 0.000587201 ENSG00000111110.10 PPM1H 0.25 8.80E-08 ENSG00000242960.1 FTH1P23 0.25 0.001099135 ENSG00000162595.4 DIRAS3 0.25 0.013273244 ENSG00000118508.4 RAB32 0.25 0.000760272 ENSG00000197852.9 FAM212B 0.25 0.00011257 ENSG00000186187.10 ZNRF1 0.25 0.010784714 ENSG00000244230.3 RN7SL151P 0.25 0.008182078 ENSG00000275379.1 HIST1H3I 0.24 0.016360538 ENSG00000133597.8 ADCK2 0.24 0.007920691 ENSG00000156299.11 TIAM1 0.24 6.28E-07 ENSG00000207513.1 RNU1-3 0.24 0.011165164 ENSG00000111640.13 GAPDH 0.24 0.000900674 ENSG00000212464.1 SNORA12 0.24 0.001354953 ENSG00000240342.3 RPS2P5 0.24 4.39E-07 ENSG00000106089.10 STX1A 0.24 0.014443212 ENSG00000120594.15 PLXDC2 0.24 3.94E-05 ENSG00000224114.1 RP11-343H5.4 0.24 0.002056595 ENSG00000131435.11 PDLIM4 0.24 3.75E-05 ENSG00000197261.10 C6orf141 0.24 0.000453799 ENSG00000006015.16 C19orf60 0.24 0.047632191 ENSG00000196372.11 ASB13 0.24 0.012156783 ENSG00000070501.10 POLB 0.24 0.003580074 ENSG00000186577.10 C6orf1 0.24 0.000720456 ENSG00000226525.5 RPS7P10 0.24 0.037777825 ENSG00000169116.10 PARM1 0.24 5.56E-07

166

ENSG00000278828.1 HIST1H3H 0.24 0.018845532 ENSG00000205403.11 CFI 0.24 6.75E-08 ENSG00000154978.11 VOPP1 0.24 5.67E-08 ENSG00000122359.16 ANXA11 0.24 1.64E-07 ENSG00000201998.1 SNORA23 0.24 0.028668277 ENSG00000173272.12 MZT2A 0.24 0.000506054 ENSG00000102871.14 TRADD 0.24 0.014799723 ENSG00000196747.4 HIST1H2AI 0.23 0.02541883 ENSG00000072210.17 ALDH3A2 0.23 5.13E-09 ENSG00000137642.11 SORL1 0.23 0.001374117 ENSG00000119655.7 NPC2 0.23 0.000862049 ENSG00000187953.9 PMS2CL 0.23 0.000119149 ENSG00000102144.12 PGK1 0.23 0.004022192 ENSG00000196878.11 LAMB3 0.23 0.001560219 ENSG00000186743.2 TPI1P3 0.23 0.000930546 ENSG00000130720.11 FIBCD1 0.23 0.016350973 ENSG00000101187.14 SLCO4A1 0.23 0.010091338 ENSG00000002822.14 MAD1L1 0.23 0.001501647 ENSG00000182621.15 PLCB1 0.23 0.002573094 ENSG00000145730.19 PAM 0.23 4.76E-05 ENSG00000278637.1 HIST1H4A 0.23 0.023100557 ENSG00000067225.16 PKM 0.23 1.03E-07 ENSG00000237172.3 B3GNT9 0.23 0.001278271 ENSG00000008300.13 CELSR3 0.23 0.048668105 ENSG00000152056.15 AP1S3 0.23 0.025497491 ENSG00000127824.12 TUBA4A 0.23 0.024112837 ENSG00000072422.15 RHOBTB1 0.23 1.22E-05 ENSG00000163710.6 PCOLCE2 0.23 0.020518618 ENSG00000146267.11 FAXC 0.23 0.003631108 ENSG00000146729.8 GBAS 0.23 4.39E-05 ENSG00000275714.1 HIST1H3A 0.23 0.031255751 ENSG00000136859.8 ANGPTL2 0.23 0.028668277 ENSG00000106009.14 BRAT1 0.23 0.004044454 ENSG00000159023.17 EPB41 0.23 0.008444711 ENSG00000105281.11 SLC1A5 0.23 0.000388846 ENSG00000269893.5 SNHG8 0.23 0.001967509 ENSG00000223803.1 RPS20P14 0.23 0.000128781 ENSG00000106266.7 SNX8 0.23 0.005303257 ENSG00000164880.14 INTS1 0.23 0.002710542 ENSG00000011638.9 TMEM159 0.23 0.012128081

167

ENSG00000164818.14 DNAAF5 0.22 0.002452139 ENSG00000118900.13 UBN1 0.22 8.14E-06 ENSG00000049541.9 RFC2 0.22 0.006019296 ENSG00000212443.1 SNORA53 0.22 0.015676951 ENSG00000197892.11 KIF13B 0.22 0.000147963 ENSG00000072682.17 P4HA2 0.22 0.029246568 ENSG00000161638.9 ITGA5 0.22 0.00030216 ENSG00000171552.11 BCL2L1 0.22 0.003540501 ENSG00000233913.7 CTC-575D19.1 0.22 3.80E-05 ENSG00000198168.7 SVIP 0.22 0.008098885 ENSG00000173801.15 JUP 0.22 0.00164837 ENSG00000132334.15 PTPRE 0.22 0.007678142 ENSG00000177181.13 RIMKLA 0.22 0.025051254 ENSG00000227939.1 RPL3P2 0.22 0.010089975 ENSG00000176597.10 B3GNT5 0.22 0.007299594 ENSG00000198355.4 PIM3 0.22 0.003055716 ENSG00000108797.10 CNTNAP1 0.22 0.027747739 ENSG00000180537.11 RNF182 0.22 1.83E-05 ENSG00000184897.5 H1FX 0.22 0.026932757 ENSG00000102048.14 ASB9 0.22 0.046018578 ENSG00000014914.18 MTMR11 0.22 0.003969911 ENSG00000169100.11 SLC25A6 0.22 3.70E-06 ENSG00000154803.11 FLCN 0.22 0.000601735 ENSG00000156521.12 TYSND1 0.22 0.006121595 ENSG00000176399.3 DMRTA1 0.22 6.89E-05 ENSG00000197956.8 S100A6 0.22 0.001041361 ENSG00000154359.11 LONRF1 0.22 0.000480575 ENSG00000101278.6 RPS10L 0.22 0.047256109 ENSG00000277157.1 HIST1H4D 0.22 0.011945588 ENSG00000135631.14 RAB11FIP5 0.22 0.002710542 ENSG00000164916.10 FOXK1 0.22 0.002477633 ENSG00000103319.10 EEF2K 0.22 0.00017654 ENSG00000197467.12 COL13A1 0.22 4.60E-05 ENSG00000196787.3 HIST1H2AG 0.22 0.046831398 ENSG00000087077.10 TRIP6 0.22 0.002522903 ENSG00000176887.6 SOX11 0.22 5.54E-06 ENSG00000168077.12 SCARA3 0.22 0.003184466 ENSG00000182253.13 SYNM 0.22 0.000569407 ENSG00000065268.9 WDR18 0.22 0.043378034 ENSG00000103152.10 MPG 0.22 0.00321226

168

ENSG00000198108.3 CHSY3 0.22 0.018945687 ENSG00000137767.12 SQRDL 0.22 0.04824413 ENSG00000170775.2 GPR37 0.22 0.001914307 ENSG00000067177.13 PHKA1 0.22 0.005274398 ENSG00000164305.16 CASP3 0.22 0.004243233 ENSG00000182372.6 CLN8 0.22 0.002764172 ENSG00000169750.7 RAC3 0.22 0.011208233 ENSG00000155893.10 PXYLP1 0.22 0.018177524 ENSG00000117691.8 NENF 0.21 0.002406869 ENSG00000111057.9 KRT18 0.21 0.000842811 ENSG00000165078.10 CPA6 0.21 0.001483458 ENSG00000172216.5 CEBPB 0.21 0.023100557 ENSG00000164796.16 CSMD3 0.21 0.017835037 ENSG00000131652.12 THOC6 0.21 0.041458921 ENSG00000213290.4 PGK1P2 0.21 0.016229784 ENSG00000136247.13 ZDHHC4 0.21 0.001112394 ENSG00000276171.1 AC114498.1 0.21 0.02837592 ENSG00000178397.11 FAM220A 0.21 0.001747617 ENSG00000244245.1 RP11-120B7.1 0.21 0.002103606 ENSG00000196843.14 ARID5A 0.21 0.000720456 ENSG00000278705.1 HIST1H4B 0.21 0.00597119 ENSG00000092964.15 DPYSL2 0.21 3.48E-05 ENSG00000152465.16 NMT2 0.21 0.005904411 ENSG00000106153.12 CHCHD2 0.21 0.000128781 ENSG00000235508.3 RPS2P7 0.21 0.00024565 ENSG00000175197.9 DDIT3 0.21 0.020297271 ENSG00000139370.9 SLC15A4 0.21 0.005994926 ENSG00000184584.11 TMEM173 0.21 0.028027511 ENSG00000162413.15 KLHL21 0.21 0.004112515 ENSG00000069702.9 TGFBR3 0.21 0.002914614 ENSG00000173511.8 VEGFB 0.21 0.014754388 ENSG00000106948.15 AKNA 0.21 0.032124885 ENSG00000141542.9 RAB40B 0.21 0.003288616 ENSG00000197256.9 KANK2 0.21 0.000279422 ENSG00000105202.6 FBL 0.21 0.010993733 ENSG00000164828.16 SUN1 0.21 1.55E-06 ENSG00000233822.4 HIST1H2BN 0.21 0.011626052 ENSG00000160703.14 NLRX1 0.21 0.02448772 ENSG00000275713.1 HIST1H2BH 0.21 0.025722531 ENSG00000005059.14 CCDC109B 0.21 0.003391966

169

ENSG00000114738.9 MAPKAPK3 0.21 0.041157824 ENSG00000227081.5 RP11-543P15.1 0.21 0.002224403 ENSG00000121297.6 TSHZ3 0.21 0.001358041 ENSG00000121057.11 AKAP1 0.21 0.000571963 ENSG00000067182.6 TNFRSF1A 0.21 0.000220202 ENSG00000159733.12 ZFYVE28 0.20 0.02502321 ENSG00000188215.8 DCUN1D3 0.20 0.01057987 ENSG00000081923.9 ATP8B1 0.20 0.007937903 ENSG00000145623.11 OSMR 0.20 0.001092702 ENSG00000147041.10 SYTL5 0.20 0.000526287 ENSG00000164713.8 BRI3 0.20 0.001317397 ENSG00000277775.1 HIST1H3F 0.20 0.009087686 ENSG00000106012.16 IQCE 0.20 0.003694391 ENSG00000124749.15 COL21A1 0.20 0.031873759 ENSG00000245910.7 SNHG6 0.20 0.002317163 ENSG00000165246.11 NLGN4Y 0.20 0.019241132 ENSG00000100321.13 SYNGR1 0.20 0.016663472 ENSG00000178209.13 PLEC 0.20 0.012563116 ENSG00000147642.15 SYBU 0.20 0.000496164 ENSG00000111266.7 DUSP16 0.20 0.000136028 ENSG00000002586.16 CD99 0.20 5.54E-06 ENSG00000154639.17 CXADR 0.20 6.25E-06 ENSG00000173083.13 HPSE 0.20 0.011130147 ENSG00000276180.1 HIST1H4I 0.20 0.010993733 ENSG00000196872.9 KIAA1211L 0.20 0.010001507 ENSG00000166173.10 LARP6 0.20 0.024385921 ENSG00000110092.3 CCND1 0.20 8.24E-05 ENSG00000232341.2 RPL4P2 0.20 0.023833751 ENSG00000196183.5 RPS2P4 0.20 0.019241132 ENSG00000189159.14 HN1 0.20 0.000128781 ENSG00000197959.12 DNM3 0.20 0.010393694 ENSG00000187244.9 BCAM 0.20 0.017456684 ENSG00000106348.15 IMPDH1 0.20 0.021870418 ENSG00000166147.12 FBN1 0.20 1.90E-05 ENSG00000137571.9 SLCO5A1 0.20 2.99E-05 ENSG00000129038.14 LOXL1 0.20 0.005129659 ENSG00000198715.10 GLMP 0.20 0.044369659 ENSG00000223509.7 RP11-632K20.7 0.20 0.043534253 ENSG00000219470.1 RP3-337H4.6 0.20 0.025497491 ENSG00000122512.13 PMS2 0.20 0.001169286

170

ENSG00000176022.4 B3GALT6 0.20 0.005548482 ENSG00000135074.14 ADAM19 0.20 5.12E-05 ENSG00000145780.7 FEM1C 0.20 0.004935775 ENSG00000171858.16 RPS21 0.20 8.13E-06 ENSG00000142627.12 EPHA2 0.20 0.002457026 ENSG00000180817.10 PPA1 0.19 0.006886481 ENSG00000105373.17 GLTSCR2 0.19 0.003464553 ENSG00000181588.16 MEX3D 0.19 0.012431167 ENSG00000164307.11 ERAP1 0.19 0.000219736 ENSG00000158406.3 HIST1H4H 0.19 0.023809924 ENSG00000166949.14 SMAD3 0.19 0.013214812 ENSG00000145390.10 USP53 0.19 0.001219442 ENSG00000107719.8 PALD1 0.19 0.011070398 ENSG00000169432.13 SCN9A 0.19 0.007412422 ENSG00000219451.3 AC005251.3 0.19 0.041619361 ENSG00000147403.15 RPL10 0.19 2.87E-06 ENSG00000182095.13 TNRC18 0.19 0.018273191 ENSG00000110047.16 EHD1 0.19 0.034053353 ENSG00000110492.14 MDK 0.19 6.69E-05 ENSG00000183840.6 GPR39 0.19 0.035856089 ENSG00000226608.3 FTLP3 0.19 0.043753273 ENSG00000178464.6 CTD- 0.19 0.001525223 2192J16.15 ENSG00000179886.5 TIGD5 0.19 0.02907165 ENSG00000175471.18 MCTP1 0.19 0.006886481 ENSG00000095794.18 CREM 0.19 0.008235495 ENSG00000228981.3 RP11-364L4.1 0.19 0.002406869 ENSG00000066629.15 EML1 0.19 0.00146729 ENSG00000213347.9 MXD3 0.19 0.031553202 ENSG00000197858.9 GPAA1 0.19 0.00306118 ENSG00000160685.12 ZBTB7B 0.19 0.027644672 ENSG00000184500.13 PROS1 0.19 3.36E-05 ENSG00000115380.17 EFEMP1 0.19 2.48E-05 ENSG00000139239.7 RPL14P1 0.19 0.003400875 ENSG00000197442.9 MAP3K5 0.19 0.009759439 ENSG00000168003.15 SLC3A2 0.19 0.00017254 ENSG00000147100.8 SLC16A2 0.19 0.012128081 ENSG00000172031.6 EPHX4 0.19 0.037733694 ENSG00000143776.17 CDC42BPA 0.19 0.000462745 ENSG00000189343.7 RPS2P46 0.19 0.000453297

171

ENSG00000226855.1 RP11-350G8.3 0.19 0.043533067 ENSG00000117152.12 RGS4 0.19 0.014894968 ENSG00000185880.11 TRIM69 0.19 7.78E-05 ENSG00000029534.18 ANK1 0.18 0.003055716 ENSG00000140511.10 HAPLN3 0.18 0.012563116 ENSG00000239246.3 RP11-464D20.2 0.18 0.033220964 ENSG00000049860.12 HEXB 0.18 0.027667067 ENSG00000197976.9 AKAP17A 0.18 0.01316159 ENSG00000185215.7 TNFAIP2 0.18 0.009339803 ENSG00000158373.8 HIST1H2BD 0.18 0.001261227 ENSG00000170017.11 ALCAM 0.18 0.00069779 ENSG00000171206.12 TRIM8 0.18 0.012174666 ENSG00000122490.17 PQLC1 0.18 0.012523054 ENSG00000272196.2 HIST2H2AA4 0.18 0.012319219 ENSG00000236439.4 RP11-175B9.3 0.18 9.57E-05 ENSG00000168209.4 DDIT4 0.18 0.01726422 ENSG00000110011.12 DNAJC4 0.18 0.033207549 ENSG00000132432.12 SEC61G 0.18 0.001704448 ENSG00000130731.14 C16orf13 0.18 0.023411883 ENSG00000013441.14 CLK1 0.18 0.023816568 ENSG00000087086.12 FTL 0.18 0.00024565 ENSG00000161013.15 MGAT4B 0.18 0.001747617 ENSG00000129103.16 SUMF2 0.18 9.84E-05 ENSG00000149823.6 VPS51 0.18 0.036912417 ENSG00000125817.7 CENPB 0.18 0.0271898 ENSG00000105976.13 MET 0.18 0.000422009 ENSG00000205903.5 ZNF316 0.18 0.009649367 ENSG00000087074.7 PPP1R15A 0.18 0.025693738 ENSG00000198242.12 RPL23A 0.18 0.003089645 ENSG00000140988.14 RPS2 0.18 0.000418983 ENSG00000068971.12 PPP2R5B 0.18 0.03521652 ENSG00000142173.13 COL6A2 0.18 0.02261305 ENSG00000169567.10 HINT1 0.18 0.019136789 ENSG00000014216.14 CAPN1 0.18 0.005370943 ENSG00000111678.9 C12orf57 0.18 0.013310609 ENSG00000219928.2 RP11-40C6.2 0.18 0.006363681 ENSG00000219545.8 UMAD1 0.18 0.046957647 ENSG00000116729.12 WLS 0.18 0.005535403 ENSG00000121578.11 B4GALT4 0.18 0.033194279 ENSG00000167526.12 RPL13 0.18 4.01E-05

172

ENSG00000124145.6 SDC4 0.18 0.003850843 ENSG00000111371.14 SLC38A1 0.18 1.18E-06 ENSG00000172890.10 NADSYN1 0.18 0.000439477 ENSG00000153922.9 CHD1 0.18 0.000109606 ENSG00000164684.12 ZNF704 0.17 0.013787589 ENSG00000178695.5 KCTD12 0.17 0.002394861 ENSG00000160959.6 LRRC14 0.17 0.025185934 ENSG00000030582.15 GRN 0.17 0.009625411 ENSG00000198890.7 PRMT6 0.17 0.023204631 ENSG00000230076.1 AC016708.2 0.17 0.003561002 ENSG00000072310.15 SREBF1 0.17 0.01747905 ENSG00000197565.14 COL4A6 0.17 0.008165025 ENSG00000120324.6 PCDHB10 0.17 0.048418365 ENSG00000177954.10 RPS27 0.17 2.80E-05 ENSG00000122870.10 BICC1 0.17 0.001738101 ENSG00000140743.6 CDR2 0.17 0.029251474 ENSG00000173548.8 SNX33 0.17 0.035539629 ENSG00000167703.13 SLC43A2 0.17 0.014161841 ENSG00000178585.13 CTNNBIP1 0.17 0.005356308 ENSG00000166741.6 NNMT 0.17 0.014434004 ENSG00000008988.8 RPS20 0.17 0.000643881 ENSG00000244731.6 C4A 0.17 0.01017216 ENSG00000101439.7 CST3 0.17 0.020023119 ENSG00000277075.1 HIST1H2AE 0.17 0.045413144 ENSG00000148082.8 SHC3 0.17 0.001170814 ENSG00000187840.4 EIF4EBP1 0.17 0.040564891 ENSG00000165175.14 MID1IP1 0.17 0.042350987 ENSG00000049239.11 H6PD 0.17 0.013117006 ENSG00000043143.19 JADE2 0.17 0.023833751 ENSG00000113441.14 LNPEP 0.17 0.000251863 ENSG00000224389.7 C4B 0.17 0.016104498 ENSG00000104964.13 AES 0.17 0.000537541 ENSG00000137818.10 RPLP1 0.17 0.001491598 ENSG00000143545.7 RAB13 0.17 0.002858719 ENSG00000077942.16 FBLN1 0.17 0.040139682 ENSG00000123144.9 C19orf43 0.17 0.010896769 ENSG00000244086.1 RPS20P35 0.17 0.048300348 ENSG00000089157.14 RPLP0 0.17 0.002115689 ENSG00000113552.14 GNPDA1 0.17 0.022722816 ENSG00000106367.12 AP1S1 0.17 0.020600867

173

ENSG00000133169.5 BEX1 0.17 0.003346493 ENSG00000189319.12 FAM53B 0.17 0.003579538 ENSG00000159840.14 ZYX 0.17 0.045096305 ENSG00000111669.13 TPI1 0.17 0.002353837 ENSG00000075711.19 DLG1 0.17 0.001686071 ENSG00000234851.4 RP11-3P17.3 0.17 0.007299594 ENSG00000122674.10 CCZ1 0.17 0.033210399 ENSG00000108846.14 ABCC3 0.17 0.002663827 ENSG00000132359.12 RAP1GAP2 0.17 0.026786903 ENSG00000143198.11 MGST3 0.17 0.028667424 ENSG00000162910.17 MRPL55 0.17 0.045682848 ENSG00000166189.7 HPS6 0.17 0.042703599 ENSG00000103855.16 CD276 0.17 0.001219442 ENSG00000240230.4 COX19 0.17 0.045096305 ENSG00000100242.14 SUN2 0.17 0.014641141 ENSG00000154640.13 BTG3 0.17 0.037435985 ENSG00000025039.13 RRAGD 0.17 0.013976642 ENSG00000184371.12 CSF1 0.16 0.017925999 ENSG00000181396.11 OGFOD3 0.16 0.010122021 ENSG00000164733.19 CTSB 0.16 0.002522903 ENSG00000165102.13 HGSNAT 0.16 0.044369659 ENSG00000155189.10 AGPAT5 0.16 0.000446867 ENSG00000163349.20 HIPK1 0.16 0.003931479 ENSG00000121957.11 GPSM2 0.16 0.002028671 ENSG00000010295.18 IFFO1 0.16 0.032428083 ENSG00000117868.14 ESYT2 0.16 0.000130052 ENSG00000138316.9 ADAMTS14 0.16 0.035375173 ENSG00000170776.18 AKAP13 0.16 0.013707249 ENSG00000157240.3 FZD1 0.16 0.013977632 ENSG00000143537.12 ADAM15 0.16 0.030398516 ENSG00000135446.15 CDK4 0.16 0.01336464 ENSG00000004059.9 ARF5 0.16 0.021112584 ENSG00000110852.4 CLEC2B 0.16 0.024274094 ENSG00000168487.16 BMP1 0.16 0.003312108 ENSG00000092621.10 PHGDH 0.16 0.018440423 ENSG00000163159.10 VPS72 0.16 0.00925877 ENSG00000197258.5 EIF4BP6 0.16 0.007678142 ENSG00000230207.1 RPL4P5 0.16 0.002245627 ENSG00000164951.14 PDP1 0.16 0.000648615 ENSG00000151474.18 FRMD4A 0.16 0.019977091

174

ENSG00000116871.14 MAP7D1 0.16 0.008203491 ENSG00000005884.16 ITGA3 0.16 0.025181686 ENSG00000183023.17 SLC8A1 0.16 0.03492123 ENSG00000185222.7 WBP5 0.16 0.017958962 ENSG00000011275.17 RNF216 0.16 0.001716853 ENSG00000229638.1 RPL4P4 0.16 0.000280177 ENSG00000157637.11 SLC38A10 0.16 0.012895306 ENSG00000154277.11 UCHL1 0.16 0.044901335 ENSG00000099875.13 MKNK2 0.16 0.029982982 ENSG00000130255.11 RPL36 0.16 0.012022547 ENSG00000101189.6 MRGBP 0.16 0.029950671 ENSG00000136261.13 BZW2 0.16 0.017066769 ENSG00000197061.4 HIST1H4C 0.16 0.042283566 ENSG00000166025.16 AMOTL1 0.16 0.005302599 ENSG00000215093.3 EEF1A1P29 0.16 0.048668105 ENSG00000060138.11 YBX3 0.16 0.002787049 ENSG00000163900.9 TMEM41A 0.16 0.029780678 ENSG00000142156.13 COL6A1 0.16 0.028081496 ENSG00000132275.9 RRP8 0.16 0.024355375 ENSG00000174444.13 RPL4 0.16 0.000128781 ENSG00000109586.10 GALNT7 0.16 0.047256109 ENSG00000026508.15 CD44 0.16 0.020521745 ENSG00000115268.8 RPS15 0.16 0.018771936 ENSG00000114861.17 FOXP1 0.15 0.024355375 ENSG00000151693.8 ASAP2 0.15 0.023411883 ENSG00000164308.15 ERAP2 0.15 0.038692414 ENSG00000006282.18 SPATA20 0.15 0.007359669 ENSG00000142534.5 RPS11 0.15 0.000627693 ENSG00000226415.1 TPI1P1 0.15 0.038026555 ENSG00000149115.12 TNKS1BP1 0.15 0.0109027 ENSG00000214717.8 ZBED1 0.15 0.005934977 ENSG00000111846.14 GCNT2 0.15 0.006621495 ENSG00000137962.11 ARHGAP29 0.15 0.005870999 ENSG00000128272.13 ATF4 0.15 0.007866219 ENSG00000111696.10 NT5DC3 0.15 0.001751147 ENSG00000237729.2 AC002075.4 0.15 0.022563678 ENSG00000136238.16 RAC1 0.15 0.014722065 ENSG00000064932.14 SBNO2 0.15 0.045096305 ENSG00000161016.14 RPL8 0.15 0.000970623 ENSG00000104783.10 KCNN4 0.15 0.020600867

175

ENSG00000184743.11 ATL3 0.15 0.006664442 ENSG00000006327.12 TNFRSF12A 0.15 0.045379119 ENSG00000123064.11 DDX54 0.15 0.028027511 ENSG00000177469.12 PTRF 0.15 0.000275057 ENSG00000213553.4 RPLP0P6 0.15 0.031936129 ENSG00000099810.17 MTAP 0.15 0.001392575 ENSG00000121022.12 COPS5 0.15 0.030186004 ENSG00000129691.14 ASH2L 0.15 0.043333473 ENSG00000162496.7 DHRS3 0.15 0.019794149 ENSG00000083845.7 RPS5 0.15 0.011853454 ENSG00000049130.12 KITLG 0.15 0.021069809 ENSG00000169255.12 B3GALNT1 0.15 0.045716568 ENSG00000240036.4 RP11-587D21.1 0.15 0.04050225 ENSG00000177606.6 JUN 0.15 0.045018328 ENSG00000139641.11 ESYT1 0.15 0.007762332 ENSG00000213614.8 HEXA 0.15 0.033892048 ENSG00000198911.10 SREBF2 0.15 0.027936577 ENSG00000107404.16 DVL1 0.15 0.036738453 ENSG00000106263.16 EIF3B 0.15 0.004351764 ENSG00000176978.12 DPP7 0.15 0.036173617 ENSG00000189043.8 NDUFA4 0.15 0.028997811 ENSG00000153113.22 CAST 0.15 0.001501647 ENSG00000212802.4 RPL15P3 0.15 0.000646801 ENSG00000188243.11 COMMD6 0.15 0.022572575 ENSG00000172059.9 KLF11 0.15 0.042166585 ENSG00000197157.9 SND1 0.15 0.011711227 ENSG00000140937.12 CDH11 0.15 0.003694391 ENSG00000135842.15 FAM129A 0.15 0.002516656 ENSG00000172809.11 RPL38 0.15 0.025497491 ENSG00000233927.4 RPS28 0.15 0.04863245 ENSG00000212232.1 SNORD17 0.15 0.047632191 ENSG00000086232.11 EIF2AK1 0.15 0.007045399 ENSG00000120708.15 TGFBI 0.15 0.021491382 ENSG00000109436.7 TBC1D9 0.15 0.003324315 ENSG00000217128.10 FNIP1 0.14 0.023280145 ENSG00000172432.17 GTPBP2 0.14 0.027593346 ENSG00000145425.8 RPS3A 0.14 0.023688246 ENSG00000102699.5 PARP4 0.14 0.006486056 ENSG00000198951.10 NAGA 0.14 0.029708337 ENSG00000170892.9 TSEN34 0.14 0.032068523

176

ENSG00000155561.13 NUP205 0.14 0.017143981 ENSG00000102178.11 UBL4A 0.14 0.037604584 ENSG00000123496.6 IL13RA2 0.14 0.024224621 ENSG00000196531.9 NACA 0.14 0.001255226 ENSG00000173898.10 SPTBN2 0.14 0.023206772 ENSG00000143333.6 RGS16 0.14 0.00907192 ENSG00000101161.7 PRPF6 0.14 0.012507548 ENSG00000205871.5 RPS3AP47 0.14 0.042350987 ENSG00000173848.17 NET1 0.14 0.011452989 ENSG00000107816.16 LZTS2 0.14 0.033207549 ENSG00000141569.9 TRIM65 0.14 0.037435985 ENSG00000140575.11 IQGAP1 0.14 0.001524388 ENSG00000176871.7 WSB2 0.14 0.017143981 ENSG00000038382.16 TRIO 0.14 0.012523054 ENSG00000204628.10 GNB2L1 0.14 0.00036032 ENSG00000108039.16 XPNPEP1 0.14 0.004836941 ENSG00000173113.5 TRMT112 0.14 0.048668105 ENSG00000104765.13 BNIP3L 0.14 0.019203978 ENSG00000150991.13 UBC 0.14 0.043534253 ENSG00000151883.15 PARP8 0.14 0.048453396 ENSG00000125691.11 RPL23 0.14 0.001208314 ENSG00000146733.12 PSPH 0.14 0.032407013 ENSG00000108861.7 DUSP3 0.14 0.005588292 ENSG00000106665.14 CLIP2 0.14 0.043869638 ENSG00000150625.15 GPM6A 0.14 0.007586448 ENSG00000196365.10 LONP1 0.14 0.032957569 ENSG00000157227.11 MMP14 0.14 0.030186004 ENSG00000111144.8 LTA4H 0.14 0.01017216 ENSG00000225031.1 EIF4BP7 0.14 0.020625865 ENSG00000149925.15 ALDOA 0.14 0.045045684 ENSG00000164631.17 ZNF12 0.14 0.016775339 ENSG00000198815.7 FOXJ3 0.14 0.036558492 ENSG00000100353.16 EIF3D 0.13 0.00568113 ENSG00000145592.12 RPL37 0.13 0.001120891 ENSG00000113580.13 NR3C1 0.13 0.003677562 ENSG00000100316.14 RPL3 0.13 0.001688331 ENSG00000120889.11 TNFRSF10B 0.13 0.027667067 ENSG00000166750.8 SLFN5 0.13 0.021935358 ENSG00000104687.11 GSR 0.13 0.049184739 ENSG00000234741.6 GAS5 0.13 0.006139275

177

ENSG00000164587.10 RPS14 0.13 0.006310552 ENSG00000182899.13 RPL35A 0.13 0.002549761 ENSG00000234004.4 RP11-543B16.1 0.13 0.046012722 ENSG00000104365.12 IKBKB 0.13 0.037110617 ENSG00000120709.9 FAM53C 0.13 0.045413144 ENSG00000238103.4 RPL9P7 0.13 0.019215401 ENSG00000187123.13 LYPD6 0.13 0.037604584 ENSG00000112893.8 MAN2A1 0.13 0.010812761 ENSG00000171863.11 RPS7 0.13 0.034127603 ENSG00000224546.2 EIF4BP3 0.13 0.02852271 ENSG00000213741.7 RPS29 0.13 0.009788603 ENSG00000215030.5 RPL13P12 0.13 0.032104503 ENSG00000100129.16 EIF3L 0.13 0.007340148 ENSG00000131469.11 RPL27 0.13 0.01726422 ENSG00000196465.9 MYL6B 0.13 0.040629686 ENSG00000136205.15 TNS3 0.13 0.032171613 ENSG00000197746.12 PSAP 0.13 0.006513539 ENSG00000144713.11 RPL32 0.13 0.005303257 ENSG00000213719.7 CLIC1 0.13 0.036358137 ENSG00000184613.9 NELL2 0.13 0.019642775 ENSG00000105612.7 DNASE2 0.13 0.038850111 ENSG00000015676.16 NUDCD3 0.13 0.04039576 ENSG00000131016.15 AKAP12 0.12 0.016076433 ENSG00000180801.12 ARSJ 0.12 0.017925999 ENSG00000149273.13 RPS3 0.12 0.005701019 ENSG00000198034.9 RPS4X 0.12 0.012523054 ENSG00000232573.1 RPL3P4 0.12 0.021901916 ENSG00000122406.11 RPL5 0.12 0.007607211 ENSG00000177600.7 RPLP2 0.12 0.012516908 ENSG00000144724.17 PTPRG 0.12 0.023100557 ENSG00000221983.6 UBA52 0.12 0.017390492 ENSG00000165502.6 RPL36AL 0.12 0.046451243 ENSG00000156515.20 HK1 0.12 0.020596079 ENSG00000188846.12 RPL14 0.12 0.008723788 ENSG00000124570.16 SERPINB6 0.12 0.032994249 ENSG00000146376.9 ARHGAP18 0.12 0.029388219 ENSG00000130741.9 EIF2S3 0.12 0.014723284 ENSG00000169760.16 NLGN1 0.12 0.046489425 ENSG00000167658.14 EEF2 0.12 0.012563116 ENSG00000147065.15 MSN 0.12 0.034643968

178

ENSG00000135390.16 ATP5G2 0.12 0.021901916 ENSG00000118181.9 RPS25 0.12 0.011987653 ENSG00000257093.5 KIAA1147 0.12 0.046018369 ENSG00000168672.3 FAM84B 0.11 0.035375173 ENSG00000149806.9 FAU 0.11 0.048724574 ENSG00000105778.16 AVL9 0.11 0.046002687 ENSG00000099622.12 CIRBP 0.11 0.019241062 ENSG00000241429.1 EEF1A1P25 0.11 0.03864122 ENSG00000108298.8 RPL19 0.11 0.029251474 ENSG00000218426.5 RP11-475C16.1 0.11 0.046707817 ENSG00000105894.10 PTN 0.11 0.014158102 ENSG00000142937.10 RPS8 0.11 0.021132706 ENSG00000142541.15 RPL13A 0.10 0.024805485 ENSG00000006468.12 ETV1 0.10 0.043256113 ENSG00000228502.1 EEF1A1P11 0.10 0.049936734 ENSG00000072778.18 ACADVL 0.10 0.042350987 ENSG00000196083.8 IL1RAP 0.10 0.04238796 ENSG00000060982.13 BCAT1 0.10 0.027037514 ENSG00000063046.16 EIF4B 0.10 0.03663808 ENSG00000198961.8 PJA2 0.09 0.045096305 ENSG00000136156.11 ITM2B -0.10 0.048379122 ENSG00000108406.8 DHX40 -0.10 0.049753087 ENSG00000044574.7 HSPA5 -0.10 0.049036676 ENSG00000156804.6 FBXO32 -0.10 0.035911445 ENSG00000135052.15 GOLM1 -0.11 0.033892048 ENSG00000035403.15 VCL -0.11 0.029814295 ENSG00000106723.15 SPIN1 -0.11 0.043333473 ENSG00000119396.9 RAB14 -0.11 0.048724574 ENSG00000065802.10 ASB1 -0.11 0.027556369 ENSG00000275052.3 SMEK2 -0.11 0.018505509 ENSG00000142166.11 IFNAR1 -0.12 0.046477696 ENSG00000010810.16 FYN -0.12 0.021591233 ENSG00000116001.14 TIA1 -0.12 0.047558821 ENSG00000164975.14 SNAPC3 -0.12 0.049748371 ENSG00000108604.14 SMARCD2 -0.12 0.037475138 ENSG00000134982.15 APC -0.12 0.017279701 ENSG00000128487.15 SPECC1 -0.12 0.019215401 ENSG00000117020.15 AKT3 -0.12 0.043417863 ENSG00000082898.15 XPO1 -0.12 0.013110789 ENSG00000062725.8 APPBP2 -0.12 0.015032202

179

ENSG00000183054.10 RGPD6 -0.12 0.013169219 ENSG00000114126.16 TFDP2 -0.12 0.01936242 ENSG00000169629.10 RGPD8 -0.12 0.011452989 ENSG00000266028.6 SRGAP2 -0.12 0.021069809 ENSG00000107290.12 SETX -0.12 0.025185934 ENSG00000015568.11 RGPD5 -0.12 0.010865279 ENSG00000066084.11 DIP2B -0.13 0.045018328 ENSG00000132300.17 PTCD3 -0.13 0.025062396 ENSG00000146072.6 TNFRSF21 -0.13 0.015432885 ENSG00000082258.11 CCNT2 -0.13 0.033162539 ENSG00000005238.18 FAM214B -0.13 0.048668105 ENSG00000114857.16 NKTR -0.13 0.035033775 ENSG00000112319.16 EYA4 -0.13 0.021475327 ENSG00000122042.9 UBL3 -0.13 0.046362128 ENSG00000090612.19 ZNF268 -0.13 0.039435238 ENSG00000036257.11 CUL3 -0.13 0.037545132 ENSG00000105810.8 CDK6 -0.13 0.031138585 ENSG00000132424.13 PNISR -0.13 0.017925999 ENSG00000241612.1 RP13-585F24.1 -0.13 0.026488987 ENSG00000176853.14 FAM91A1 -0.13 0.02502321 ENSG00000170456.13 DENND5B -0.13 0.020581919 ENSG00000270872.2 SRGAP2D -0.13 0.030019387 ENSG00000183853.16 KIRREL -0.14 0.027747739 ENSG00000253352.7 TUG1 -0.14 0.013283786 ENSG00000075292.17 ZNF638 -0.14 0.005554107 ENSG00000140262.16 TCF12 -0.14 0.001867521 ENSG00000100852.11 ARHGAP5 -0.14 0.01679754 ENSG00000170558.7 CDH2 -0.14 0.003324315 ENSG00000143797.10 MBOAT2 -0.14 0.005847074 ENSG00000135968.18 GCC2 -0.14 0.028074567 ENSG00000094975.12 SUCO -0.14 0.003744904 ENSG00000107362.12 ABHD17B -0.14 0.03663808 ENSG00000070214.14 SLC44A1 -0.14 0.026644688 ENSG00000145632.13 PLK2 -0.14 0.017184787 ENSG00000167552.12 TUBA1A -0.14 0.024805485 ENSG00000181826.8 RELL1 -0.14 0.020462756 ENSG00000220842.6 RP11-572P18.1 -0.14 0.004428565 ENSG00000166479.8 TMX3 -0.14 0.042018672 ENSG00000147526.18 TACC1 -0.14 0.00142086 ENSG00000137942.15 FNBP1L -0.14 0.001000446

180

ENSG00000220749.4 RPL21P28 -0.14 0.009324879 ENSG00000171492.13 LRRC8D -0.14 0.00770033 ENSG00000220793.5 RPL21P119 -0.14 0.00874526 ENSG00000164347.16 GFM2 -0.14 0.021475327 ENSG00000169155.8 ZBTB43 -0.14 0.038765873 ENSG00000095015.5 MAP3K1 -0.15 0.019125741 ENSG00000136478.6 TEX2 -0.15 0.011399224 ENSG00000159086.13 PAXBP1 -0.15 0.036897448 ENSG00000180957.16 PITPNB -0.15 0.045192664 ENSG00000180628.13 PCGF5 -0.15 0.019683452 ENSG00000156671.11 SAMD8 -0.15 0.004836941 ENSG00000109790.15 KLHL5 -0.15 0.039771855 ENSG00000235558.3 RP11- -0.15 0.037777825 1198D22.1 ENSG00000026652.12 AGPAT4 -0.15 0.023802931 ENSG00000164930.10 FZD6 -0.15 0.002544043 ENSG00000075151.18 EIF4G3 -0.15 0.00719664 ENSG00000136720.6 HS6ST1 -0.15 0.0020095 ENSG00000044524.9 EPHA3 -0.15 0.000285842 ENSG00000139737.20 SLAIN1 -0.15 0.033601764 ENSG00000019144.15 PHLDB1 -0.15 0.011491626 ENSG00000163513.16 TGFBR2 -0.15 0.003111058 ENSG00000196628.12 TCF4 -0.15 0.000523577 ENSG00000254004.5 ZNF260 -0.15 0.014641141 ENSG00000143013.11 LMO4 -0.15 0.002171677 ENSG00000253797.2 UTP14C -0.15 0.029646542 ENSG00000223705.8 NSUN5P1 -0.15 0.019136789 ENSG00000119414.10 PPP6C -0.15 0.029246568 ENSG00000196302.5 RP11-497H16.5 -0.15 0.0324087 ENSG00000146676.6 PURB -0.15 0.016989761 ENSG00000182903.14 ZNF721 -0.15 0.027225667 ENSG00000076604.13 TRAF4 -0.15 0.010436263 ENSG00000196369.9 SRGAP2B -0.15 0.000767314 ENSG00000143324.12 XPR1 -0.15 0.024355375 ENSG00000122026.9 RPL21 -0.15 0.001818338 ENSG00000064042.16 LIMCH1 -0.15 0.006257388 ENSG00000082996.18 RNF13 -0.15 0.02381396 ENSG00000274602.3 PI4KAP1 -0.15 0.045846394 ENSG00000111817.15 DSE -0.16 0.028668277 ENSG00000183506.15 PI4KAP2 -0.16 0.021935358

181

ENSG00000011454.15 RABGAP1 -0.16 0.002755746 ENSG00000073712.12 FERMT2 -0.16 0.008640701 ENSG00000153823.17 PID1 -0.16 0.010297192 ENSG00000135414.8 GDF11 -0.16 0.017720961 ENSG00000173402.10 DAG1 -0.16 0.004145837 ENSG00000096093.13 EFHC1 -0.16 0.003579538 ENSG00000162616.8 DNAJB4 -0.16 0.030751508 ENSG00000023608.4 SNAPC1 -0.16 0.045915337 ENSG00000167232.12 ZNF91 -0.16 0.004899737 ENSG00000105767.2 CADM4 -0.16 0.012535822 ENSG00000170581.12 STAT2 -0.16 0.000110374 ENSG00000272841.1 RP3-428L16.2 -0.16 0.01739134 ENSG00000180008.8 SOCS4 -0.16 0.019538445 ENSG00000111785.17 RIC8B -0.16 0.020600867 ENSG00000164543.5 STK17A -0.16 0.020286626 ENSG00000181751.8 C5orf30 -0.16 0.047500111 ENSG00000169862.17 CTNND2 -0.16 0.017524241 ENSG00000062485.17 CS -0.16 0.001161833 ENSG00000171943.10 SRGAP2C -0.17 7.04E-05 ENSG00000135535.13 CD164 -0.17 0.000640702 ENSG00000170500.11 LONRF2 -0.17 0.048916098 ENSG00000117450.12 PRDX1 -0.17 0.018608901 ENSG00000162512.14 SDC3 -0.17 0.003008478 ENSG00000139679.14 LPAR6 -0.17 0.035424309 ENSG00000082269.15 FAM135A -0.17 0.007607211 ENSG00000113361.11 CDH6 -0.17 4.76E-05 ENSG00000132256.17 TRIM5 -0.17 0.030555682 ENSG00000117569.17 PTBP2 -0.17 0.018363563 ENSG00000251562.6 MALAT1 -0.17 0.019136789 ENSG00000100644.15 HIF1A -0.17 0.007611332 ENSG00000224078.11 SNHG14 -0.17 0.01716184 ENSG00000276291.3 RP11-87H9.2 -0.17 0.03505879 ENSG00000056097.14 ZFR -0.17 2.15E-05 ENSG00000124193.13 SRSF6 -0.17 0.000773628 ENSG00000162664.15 ZNF326 -0.17 0.005301528 ENSG00000183722.7 LHFP -0.17 0.018714275 ENSG00000165061.13 ZMAT4 -0.17 0.032104503 ENSG00000110841.12 PPFIBP1 -0.17 0.015030507 ENSG00000111790.12 FGFR1OP2 -0.17 0.003161004 ENSG00000215193.11 PEX26 -0.17 0.02307271

182

ENSG00000107186.15 MPDZ -0.17 3.73E-05 ENSG00000175066.14 GK5 -0.17 0.033207549 ENSG00000083123.13 BCKDHB -0.17 0.012523054 ENSG00000179241.11 LDLRAD3 -0.17 0.000505888 ENSG00000196705.7 ZNF431 -0.17 0.012151305 ENSG00000138131.3 LOXL4 -0.17 0.040139682 ENSG00000123636.16 BAZ2B -0.17 0.003017204 ENSG00000164904.14 ALDH7A1 -0.17 0.001686796 ENSG00000139044.9 B4GALNT3 -0.17 0.024762148 ENSG00000166532.14 RIMKLB -0.18 0.00070861 ENSG00000180626.9 ZNF594 -0.18 0.00966174 ENSG00000136866.12 ZFP37 -0.18 0.043690349 ENSG00000170802.14 FOXN2 -0.18 0.014119626 ENSG00000105875.12 WDR91 -0.18 0.004542566 ENSG00000112769.17 LAMA4 -0.18 0.016743065 ENSG00000188647.11 PTAR1 -0.18 9.72E-05 ENSG00000080298.14 RFX3 -0.18 0.012362724 ENSG00000213860.4 RPL21P75 -0.18 0.009393161 ENSG00000151576.9 QTRTD1 -0.18 0.024974446 ENSG00000197020.9 ZNF100 -0.18 0.023199805 ENSG00000175899.13 A2M -0.18 0.016989761 ENSG00000079482.12 OPHN1 -0.18 0.000246055 ENSG00000244021.4 RP11-50D9.1 -0.18 0.034605862 ENSG00000087053.17 MTMR2 -0.18 0.002007748 ENSG00000206149.9 HERC2P9 -0.18 0.004478844 ENSG00000105854.11 PON2 -0.18 0.000789165 ENSG00000276550.3 HERC2P2 -0.18 0.002916243 ENSG00000277443.1 MARCKS -0.18 0.00027301 ENSG00000187231.12 SESTD1 -0.18 0.001734285 ENSG00000180357.8 ZNF609 -0.18 0.04863245 ENSG00000064313.10 TAF2 -0.18 0.001005732 ENSG00000147874.9 HAUS6 -0.18 0.001768593 ENSG00000171766.14 GATM -0.18 0.008441511 ENSG00000151276.22 MAGI1 -0.18 0.009136202 ENSG00000159200.16 RCAN1 -0.18 0.000160334 ENSG00000179314.12 WSCD1 -0.18 0.0109027 ENSG00000055332.15 EIF2AK2 -0.18 0.018481373 ENSG00000154265.14 ABCA5 -0.18 0.017767069 ENSG00000239470.3 RP11-16F15.2 -0.18 0.019152403 ENSG00000103460.15 TOX3 -0.19 0.024270293

183

ENSG00000123595.6 RAB9A -0.19 0.005904411 ENSG00000074657.12 ZNF532 -0.19 0.00847145 ENSG00000123124.12 WWP1 -0.19 0.035772627 ENSG00000115392.10 FANCL -0.19 0.00618984 ENSG00000196776.13 CD47 -0.19 0.0002929 ENSG00000101040.18 ZMYND8 -0.19 0.004862592 ENSG00000135916.14 ITM2C -0.19 7.60E-05 ENSG00000173041.10 ZNF680 -0.19 0.019152403 ENSG00000181450.16 ZNF678 -0.19 0.014485086 ENSG00000139289.12 PHLDA1 -0.19 0.042463877 ENSG00000153208.15 MERTK -0.19 0.023100557 ENSG00000139211.6 AMIGO2 -0.19 0.003002014 ENSG00000013523.8 ANGEL1 -0.19 0.029659643 ENSG00000119314.14 PTBP3 -0.19 0.000789481 ENSG00000165156.13 ZHX1 -0.19 0.003694391 ENSG00000198538.9 ZNF28 -0.19 0.006424515 ENSG00000054965.9 FAM168A -0.19 0.013310609 ENSG00000213080.3 RP11-312J18.5 -0.19 1.87E-05 ENSG00000092847.9 AGO1 -0.19 0.012632718 ENSG00000185432.11 METTL7A -0.19 0.003808221 ENSG00000142197.11 DOPEY2 -0.19 0.04105657 ENSG00000122707.10 RECK -0.19 0.001333786 ENSG00000113594.8 LIFR -0.19 0.011130147 ENSG00000064102.13 ASUN -0.19 0.003561435 ENSG00000175595.13 ERCC4 -0.20 0.010089975 ENSG00000134986.12 NREP -0.20 0.02315123 ENSG00000198237.8 RP11-98J23.2 -0.20 0.006008249 ENSG00000112851.13 ERBB2IP -0.20 4.12E-06 ENSG00000135821.15 GLUL -0.20 2.28E-07 ENSG00000230869.1 CTGLF10P -0.20 0.029246568 ENSG00000162407.8 PPAP2B -0.20 0.033892048 ENSG00000253366.3 RP11-589F5.4 -0.20 0.004964627 ENSG00000158352.14 SHROOM4 -0.20 0.011463583 ENSG00000112419.13 PHACTR2 -0.20 0.013310609 ENSG00000088854.12 C20orf194 -0.20 0.003184466 ENSG00000175893.10 ZDHHC21 -0.20 0.000109606 ENSG00000162971.9 TYW5 -0.20 0.01605613 ENSG00000140526.15 ABHD2 -0.20 0.000524584 ENSG00000205659.9 LIN52 -0.20 0.031518456 ENSG00000196923.12 PDLIM7 -0.20 0.004158857

184

ENSG00000144283.20 PKP4 -0.20 0.007409833 ENSG00000075407.16 ZNF37A -0.21 0.003161004 ENSG00000265018.5 CTGLF12P -0.21 0.020814444 ENSG00000164983.7 TMEM65 -0.21 8.50E-05 ENSG00000247828.6 TMEM161B- -0.21 0.001194686 AS1 ENSG00000196781.12 TLE1 -0.21 2.09E-05 ENSG00000169302.13 STK32A -0.21 0.031216187 ENSG00000196739.13 COL27A1 -0.21 0.023726483 ENSG00000033170.15 FUT8 -0.21 2.28E-05 ENSG00000125945.13 ZNF436 -0.21 4.07E-06 ENSG00000153179.10 RASSF3 -0.21 0.000622844 ENSG00000255690.2 TRIL -0.21 0.010988951 ENSG00000113387.10 SUB1 -0.21 0.001015311 ENSG00000113140.9 SPARC -0.21 2.08E-07 ENSG00000152578.11 GRIA4 -0.21 0.000602659 ENSG00000188342.10 GTF2F2 -0.21 0.033099008 ENSG00000096433.9 ITPR3 -0.21 0.012174666 ENSG00000251158.1 RP11-98J23.1 -0.21 0.037139183 ENSG00000165675.15 ENOX2 -0.21 0.008531803 ENSG00000254701.3 RP11- -0.21 0.002826659 1415C14.4 ENSG00000101265.14 RASSF2 -0.22 7.11E-07 ENSG00000148143.11 ZNF462 -0.22 0.016422257 ENSG00000139668.8 WDFY2 -0.22 0.000505888 ENSG00000178809.10 TRIM73 -0.22 0.049916244 ENSG00000171310.9 CHST11 -0.22 0.00075177 ENSG00000124126.12 PREX1 -0.22 0.001362793 ENSG00000093144.17 ECHDC1 -0.22 0.014499965 ENSG00000128000.14 ZNF780B -0.22 0.001663178 ENSG00000155760.2 FZD7 -0.22 3.65E-06 ENSG00000145555.13 MYO10 -0.22 1.01E-05 ENSG00000182141.8 ZNF708 -0.22 0.001560219 ENSG00000183666.15 GUSBP1 -0.22 0.005828313 ENSG00000110721.10 CHKA -0.22 0.036878969 ENSG00000187498.13 COL4A1 -0.22 2.83E-05 ENSG00000168734.12 PKIG -0.22 0.000141274 ENSG00000102172.14 SMS -0.22 2.97E-08 ENSG00000159348.11 CYB5R1 -0.22 4.76E-05 ENSG00000078699.20 CBFA2T2 -0.22 0.008563489 ENSG00000123983.12 ACSL3 -0.22 0.006126787

185

ENSG00000009694.12 TENM1 -0.22 0.012156783 ENSG00000067445.19 TRO -0.22 0.000150007 ENSG00000139971.14 C14orf37 -0.22 0.044409254 ENSG00000182132.11 KCNIP1 -0.22 0.018781824 ENSG00000184949.14 FAM227A -0.22 0.023825837 ENSG00000205269.5 TMEM170B -0.22 2.67E-05 ENSG00000152377.11 SPOCK1 -0.23 0.000162665 ENSG00000260917.1 RP11-57H14.4 -0.23 0.036987385 ENSG00000275342.3 SGK223 -0.23 0.025307111 ENSG00000111859.15 NEDD9 -0.23 7.31E-07 ENSG00000085377.12 PREP -0.23 4.39E-05 ENSG00000170677.5 SOCS6 -0.23 3.03E-05 ENSG00000189423.10 USP32P3 -0.23 0.010578885 ENSG00000153982.9 GDPD1 -0.23 0.032762708 ENSG00000076641.4 PAG1 -0.23 0.011320416 ENSG00000253203.5 GUSBP3 -0.23 0.001097687 ENSG00000104870.11 FCGRT -0.23 0.001839883 ENSG00000196376.9 SLC35F1 -0.23 0.001747617 ENSG00000253816.3 RP11- -0.23 0.001060291 1415C14.3 ENSG00000106780.8 MEGF9 -0.23 0.004158857 ENSG00000204172.10 AGAP10 -0.23 0.007866219 ENSG00000079215.12 SLC1A3 -0.23 6.76E-05 ENSG00000198947.13 DMD -0.23 0.043154457 ENSG00000071246.9 VASH1 -0.23 0.001121775 ENSG00000095951.15 HIVEP1 -0.23 0.001861942 ENSG00000142089.14 IFITM3 -0.23 0.021901916 ENSG00000127418.13 FGFRL1 -0.23 4.56E-05 ENSG00000258441.1 LINC00641 -0.23 0.013767905 ENSG00000116525.12 TRIM62 -0.24 0.034716874 ENSG00000198729.4 PPP1R14C -0.24 0.022125601 ENSG00000169439.10 SDC2 -0.24 2.44E-05 ENSG00000174405.12 LIG4 -0.24 0.002758876 ENSG00000128606.11 LRRC17 -0.24 0.039735677 ENSG00000021645.16 NRXN3 -0.24 0.004928388 ENSG00000173517.9 PEAK1 -0.24 0.000532044 ENSG00000166348.16 USP54 -0.24 0.00159199 ENSG00000178662.14 CSRNP3 -0.24 0.004158857 ENSG00000154447.13 SH3RF1 -0.24 0.019771515 ENSG00000261069.3 RP11-701H24.4 -0.24 0.006944448

186

ENSG00000081913.12 PHLPP1 -0.24 0.001170814 ENSG00000165185.13 KIAA1958 -0.24 0.022090417 ENSG00000159256.11 MORC3 -0.24 2.04E-07 ENSG00000124615.16 MOCS1 -0.25 0.045135842 ENSG00000157657.13 ZNF618 -0.25 0.039285014 ENSG00000135069.12 PSAT1 -0.25 0.000130527 ENSG00000159921.13 GNE -0.25 3.10E-06 ENSG00000122778.8 KIAA1549 -0.25 0.017666649 ENSG00000153094.20 BCL2L11 -0.25 0.005595065 ENSG00000135926.11 TMBIM1 -0.25 0.048951452 ENSG00000153317.13 ASAP1 -0.25 0.00538066 ENSG00000006118.13 TMEM132A -0.25 1.54E-05 ENSG00000185565.10 LSAMP -0.25 0.00074153 ENSG00000233178.6 RP11-88I18.2 -0.25 0.043999203 ENSG00000144847.11 IGSF11 -0.25 0.003081887 ENSG00000229018.5 RP11-313P13.3 -0.25 0.02793211 ENSG00000080493.12 SLC4A4 -0.26 7.46E-06 ENSG00000152527.12 PLEKHH2 -0.26 0.001337741 ENSG00000105559.10 PLEKHA4 -0.26 0.001501647 ENSG00000106829.17 TLE4 -0.26 8.10E-05 ENSG00000079691.16 LRRC16A -0.26 0.001170814 ENSG00000167693.15 NXN -0.26 0.001291844 ENSG00000198521.10 ZNF43 -0.26 9.66E-06 ENSG00000009413.14 REV3L -0.26 8.49E-10 ENSG00000132122.10 SPATA6 -0.26 4.34E-06 ENSG00000105855.8 ITGB8 -0.26 3.94E-08 ENSG00000235554.1 AC005822.1 -0.26 0.004112275 ENSG00000227640.2 SOX21-AS1 -0.26 3.48E-05 ENSG00000187239.15 FNBP1 -0.26 0.000101252 ENSG00000163171.7 CDC42EP3 -0.26 1.54E-06 ENSG00000106853.15 PTGR1 -0.26 0.017945397 ENSG00000241549.7 GUSBP2 -0.26 0.000412913 ENSG00000197329.10 PELI1 -0.26 5.77E-05 ENSG00000159917.13 ZNF235 -0.27 0.045096305 ENSG00000156103.14 MMP16 -0.27 3.12E-07 ENSG00000136378.13 ADAMTS7 -0.27 0.035375173 ENSG00000259820.1 AC083843.1 -0.27 6.44E-05 ENSG00000271425.5 NBPF10 -0.27 0.005364574 ENSG00000111913.14 FAM65B -0.27 2.83E-08 ENSG00000196268.10 ZNF493 -0.27 0.000762653

187

ENSG00000117707.14 PROX1 -0.27 0.017184787 ENSG00000271533.1 RP3-368A4.6 -0.27 0.00350316 ENSG00000117643.13 MAN1C1 -0.27 0.000752557 ENSG00000154511.10 FAM69A -0.27 2.31E-07 ENSG00000126785.11 RHOJ -0.27 1.07E-08 ENSG00000112531.15 QKI -0.27 8.32E-14 ENSG00000215158.8 RP11- -0.27 0.001215305 1023L17.1 ENSG00000136158.9 SPRY2 -0.27 6.25E-13 ENSG00000138434.15 SSFA2 -0.27 3.70E-11 ENSG00000187678.8 SPRY4 -0.28 1.74E-08 ENSG00000148175.11 STOM -0.28 0.0001662 ENSG00000244582.2 RPL21P120 -0.28 0.018490386 ENSG00000136161.11 RCBTB2 -0.28 0.00193271 ENSG00000228623.3 ZNF883 -0.28 0.015411521 ENSG00000175130.6 MARCKSL1 -0.28 2.20E-07 ENSG00000213514.2 RP11-428P16.2 -0.28 0.042724856 ENSG00000236404.7 VLDLR-AS1 -0.28 0.043154457 ENSG00000216775.2 RP1-152L7.5 -0.28 0.00379563 ENSG00000170962.11 PDGFD -0.28 3.02E-11 ENSG00000114805.15 PLCH1 -0.29 0.005694652 ENSG00000182752.9 PAPPA -0.29 0.027593346 ENSG00000155324.8 GRAMD3 -0.29 3.27E-07 ENSG00000139116.16 KIF21A -0.29 1.28E-07 ENSG00000115556.12 PLCD4 -0.29 0.0005554 ENSG00000114019.13 AMOTL2 -0.29 5.67E-08 ENSG00000056998.17 GYG2 -0.29 0.001199934 ENSG00000161328.10 LRRC56 -0.29 0.004792408 ENSG00000101680.12 LAMA1 -0.29 7.90E-10 ENSG00000054690.12 PLEKHH1 -0.29 0.002249369 ENSG00000197124.10 ZNF682 -0.29 0.006424515 ENSG00000143494.14 VASH2 -0.29 0.018079709 ENSG00000146197.8 SCUBE3 -0.29 0.02494287 ENSG00000187187.12 ZNF546 -0.29 0.048087537 ENSG00000153904.17 DDAH1 -0.29 1.45E-09 ENSG00000246334.2 PRR7-AS1 -0.29 0.039071774 ENSG00000215386.9 MIR99AHG -0.30 0.000789481 ENSG00000120278.13 PLEKHG1 -0.30 3.30E-05 ENSG00000154898.14 CCDC144CP -0.30 4.47E-07 ENSG00000145284.10 SCD5 -0.30 8.05E-08

188

ENSG00000176533.11 GNG7 -0.30 5.47E-05 ENSG00000106278.10 PTPRZ1 -0.30 8.99E-10 ENSG00000184384.12 MAML2 -0.30 0.001038576 ENSG00000278416.1 PMS2L2 -0.30 0.01741473 ENSG00000135048.12 TMEM2 -0.30 5.25E-14 ENSG00000215146.4 RP11-313J2.1 -0.30 0.035989931 ENSG00000162849.14 KIF26B -0.30 0.030751508 ENSG00000215156.5 RP11- -0.30 0.007777885 1023L17.2 ENSG00000138738.9 PRDM5 -0.30 0.003697053 ENSG00000173376.12 NDNF -0.30 0.019324842 ENSG00000160781.14 PAQR6 -0.30 1.58E-05 ENSG00000101144.11 BMP7 -0.30 0.000638199 ENSG00000150551.10 LYPD1 -0.31 0.002370972 ENSG00000064692.17 SNCAIP -0.31 8.80E-08 ENSG00000111728.9 ST8SIA1 -0.31 0.000175133 ENSG00000157890.16 MEGF11 -0.31 0.001573185 ENSG00000164574.14 GALNT10 -0.31 2.36E-09 ENSG00000133026.11 MYH10 -0.31 8.50E-05 ENSG00000163536.11 SERPINI1 -0.31 0.000740566 ENSG00000145703.14 IQGAP2 -0.31 2.34E-06 ENSG00000166562.7 SEC11C -0.31 0.000379328 ENSG00000173786.15 CNP -0.31 3.68E-12 ENSG00000138035.13 PNPT1 -0.31 0.0005554 ENSG00000137628.15 DDX60 -0.31 0.026209848 ENSG00000196172.9 ZNF681 -0.31 1.78E-05 ENSG00000129657.13 SEC14L1 -0.31 1.58E-07 ENSG00000183850.12 ZNF730 -0.31 0.02631377 ENSG00000224189.5 HAGLR -0.31 0.046012722 ENSG00000188738.12 FSIP2 -0.31 0.013310609 ENSG00000250067.10 YJEFN3 -0.31 0.003636049 ENSG00000151835.12 SACS -0.31 7.07E-10 ENSG00000084710.12 EFR3B -0.31 2.14E-07 ENSG00000168389.16 MFSD2A -0.31 0.002041863 ENSG00000214595.10 EML6 -0.31 0.030226569 ENSG00000138835.21 RGS3 -0.31 5.00E-12 ENSG00000137727.11 ARHGAP20 -0.31 0.003778819 ENSG00000168016.12 TRANK1 -0.32 0.045018328 ENSG00000257354.2 RP11-631N16.2 -0.32 0.019152403 ENSG00000121005.7 CRISPLD1 -0.32 3.23E-13

189

ENSG00000133943.19 C14orf159 -0.32 0.001109051 ENSG00000062038.12 CDH3 -0.32 4.27E-06 ENSG00000124406.15 ATP8A1 -0.32 0.007607211 ENSG00000142871.14 CYR61 -0.32 0.000870294 ENSG00000246859.2 STARD4-AS1 -0.32 0.035375173 ENSG00000109819.7 PPARGC1A -0.32 0.012037663 ENSG00000130147.14 SH3BP4 -0.32 9.33E-11 ENSG00000113389.14 NPR3 -0.32 2.05E-05 ENSG00000174721.9 FGFBP3 -0.32 0.005064465 ENSG00000279555.1 RP11-426D19.1 -0.32 0.043533067 ENSG00000256771.2 ZNF253 -0.32 0.027747739 ENSG00000250420.7 AACSP1 -0.32 0.000158363 ENSG00000129675.14 ARHGEF6 -0.33 8.04E-05 ENSG00000114200.8 BCHE -0.33 1.94E-09 ENSG00000257151.1 PWAR6 -0.33 0.000720456 ENSG00000121989.13 ACVR2A -0.33 9.28E-07 ENSG00000135324.5 MRAP2 -0.33 3.45E-07 ENSG00000158186.11 MRAS -0.33 1.90E-05 ENSG00000144834.11 TAGLN3 -0.33 0.000223427 ENSG00000073464.10 CLCN4 -0.33 0.011901374 ENSG00000071282.10 LMCD1 -0.33 1.22E-06 ENSG00000106236.3 NPTX2 -0.33 0.045682848 ENSG00000116106.10 EPHA4 -0.33 1.08E-10 ENSG00000128594.6 LRRC4 -0.34 0.00907264 ENSG00000135097.5 MSI1 -0.34 0.000720456 ENSG00000159445.11 THEM4 -0.34 0.004230337 ENSG00000122378.12 FAM213A -0.34 0.000168935 ENSG00000256463.7 SALL3 -0.34 0.000345622 ENSG00000154874.13 CCDC144B -0.34 8.39E-06 ENSG00000204128.5 C2orf72 -0.34 0.031819861 ENSG00000114251.12 WNT5A -0.34 0.000119875 ENSG00000164434.10 FABP7 -0.34 4.17E-16 ENSG00000196730.11 DAPK1 -0.34 6.96E-06 ENSG00000277778.1 PGM5P2 -0.34 0.003853715 ENSG00000164056.9 SPRY1 -0.34 5.35E-13 ENSG00000163762.5 TM4SF18 -0.34 0.013565174 ENSG00000162614.17 NEXN -0.35 0.002611666 ENSG00000137936.15 BCAR3 -0.35 4.87E-12 ENSG00000154930.13 ACSS1 -0.35 2.36E-10 ENSG00000198838.10 RYR3 -0.35 0.000393711

190

ENSG00000183508.4 FAM46C -0.35 0.04863245 ENSG00000085117.10 CD82 -0.35 0.045192664 ENSG00000165323.14 FAT3 -0.35 1.10E-08 ENSG00000139926.14 FRMD6 -0.35 3.69E-14 ENSG00000185201.15 IFITM2 -0.35 0.0005554 ENSG00000127863.14 TNFRSF19 -0.35 5.23E-15 ENSG00000168461.11 RAB31 -0.36 2.00E-12 ENSG00000121068.12 TBX2 -0.36 6.67E-09 ENSG00000163638.12 ADAMTS9 -0.36 2.94E-10 ENSG00000046653.13 GPM6B -0.36 2.59E-21 ENSG00000155792.8 DEPTOR -0.36 1.32E-09 ENSG00000038427.14 VCAN -0.36 1.55E-14 ENSG00000065308.4 TRAM2 -0.36 1.20E-12 ENSG00000128849.10 CGNL1 -0.36 1.11E-15 ENSG00000258754.6 LINC01579 -0.37 0.000485466 ENSG00000198121.12 LPAR1 -0.37 1.74E-05 ENSG00000117318.8 ID3 -0.37 1.08E-10 ENSG00000248905.7 FMN1 -0.37 0.005528085 ENSG00000250111.3 AC107982.4 -0.37 0.027747739 ENSG00000163449.9 TMEM169 -0.37 4.10E-06 ENSG00000138606.18 SHF -0.37 3.05E-05 ENSG00000076356.6 PLXNA2 -0.37 4.51E-06 ENSG00000114757.17 PEX5L -0.38 0.010150396 ENSG00000103187.7 COTL1 -0.38 1.54E-07 ENSG00000258548.4 LINC00645 -0.38 0.012632718 ENSG00000170160.15 CCDC144A -0.38 9.02E-07 ENSG00000184226.13 PCDH9 -0.38 3.20E-10 ENSG00000164683.15 HEY1 -0.38 9.33E-11 ENSG00000254343.2 RP11-760H22.2 -0.38 0.011963296 ENSG00000173404.4 INSM1 -0.38 8.50E-05 ENSG00000239887.4 C1orf226 -0.38 0.000291841 ENSG00000236581.7 STARD13-AS -0.38 0.000788659 ENSG00000145824.11 CXCL14 -0.38 3.59E-06 ENSG00000095587.8 TLL2 -0.38 0.013110789 ENSG00000239374.1 RP11-407P2.1 -0.38 0.032398957 ENSG00000112276.12 BVES -0.39 2.01E-20 ENSG00000144642.19 RBMS3 -0.39 1.60E-07 ENSG00000172164.12 SNTB1 -0.39 1.54E-14 ENSG00000087253.10 LPCAT2 -0.39 4.55E-07 ENSG00000138623.8 SEMA7A -0.39 0.000586844

191

ENSG00000143341.10 HMCN1 -0.39 3.94E-08 ENSG00000148541.11 FAM13C -0.39 0.010988951 ENSG00000163235.14 TGFA -0.39 1.44E-09 ENSG00000135835.9 KIAA1614 -0.39 0.000308809 ENSG00000121871.3 SLITRK3 -0.39 0.027667067 ENSG00000163485.14 ADORA1 -0.40 0.001956869 ENSG00000070731.8 ST6GALNAC2 -0.40 0.000586239 ENSG00000136603.12 SKIL -0.40 1.58E-09 ENSG00000197013.8 ZNF429 -0.40 1.18E-06 ENSG00000238113.5 LINC01410 -0.40 0.007567886 ENSG00000168280.15 KIF5C -0.40 5.03E-19 ENSG00000092969.10 TGFB2 -0.40 0.000110628 ENSG00000147509.12 RGS20 -0.40 8.81E-07 ENSG00000179761.10 PIPOX -0.40 8.67E-14 ENSG00000154262.11 ABCA6 -0.41 0.029637666 ENSG00000146950.11 SHROOM2 -0.41 2.77E-07 ENSG00000170961.6 HAS2 -0.41 0.000141274 ENSG00000110042.6 DTX4 -0.41 2.72E-08 ENSG00000065320.7 NTN1 -0.41 2.50E-11 ENSG00000083067.21 TRPM3 -0.41 3.26E-15 ENSG00000239268.2 RP11-384F7.2 -0.41 0.003585015 ENSG00000151892.13 GFRA1 -0.42 1.46E-13 ENSG00000146070.15 PLA2G7 -0.42 0.032394139 ENSG00000087245.11 MMP2 -0.42 4.78E-17 ENSG00000114646.8 CSPG5 -0.42 0.001041361 ENSG00000170390.13 DCLK2 -0.42 6.37E-09 ENSG00000279281.1 RP11- -0.42 0.042856969 226M10.3 ENSG00000279427.1 RP11-132F7.2 -0.42 0.01810909 ENSG00000263146.2 RP11-849I19.1 -0.43 0.000154452 ENSG00000100784.8 RPS6KA5 -0.43 1.89E-06 ENSG00000145990.9 GFOD1 -0.43 0.000422009 ENSG00000229127.1 AC007038.7 -0.43 0.034482482 ENSG00000213694.3 S1PR3 -0.43 0.000418983 ENSG00000144857.13 BOC -0.43 2.81E-17 ENSG00000174136.10 RGMB -0.43 6.23E-10 ENSG00000254187.1 CTB-78F1.1 -0.44 0.035994491 ENSG00000103485.16 QPRT -0.44 3.31E-05 ENSG00000132854.17 KANK4 -0.44 0.026209848 ENSG00000138642.13 HERC6 -0.44 0.029774389

192

ENSG00000099204.17 ABLIM1 -0.44 1.23E-06 ENSG00000198929.11 NOS1AP -0.45 0.001313819 ENSG00000136002.15 ARHGEF4 -0.45 0.000143389 ENSG00000166106.3 ADAMTS15 -0.45 0.001574831 ENSG00000228213.4 NLGN1-AS1 -0.45 0.047558821 ENSG00000139132.13 FGD4 -0.45 1.62E-08 ENSG00000180113.14 TDRD6 -0.45 0.020888502 ENSG00000128567.15 PODXL -0.46 1.29E-12 ENSG00000123104.10 ITPR2 -0.46 2.01E-20 ENSG00000206557.5 TRIM71 -0.47 0.016955451 ENSG00000184635.12 ZNF93 -0.47 6.62E-05 ENSG00000083290.18 ULK2 -0.47 1.21E-05 ENSG00000141314.11 RHBDL3 -0.47 0.01918895 ENSG00000260664.2 AC004158.3 -0.47 0.029796102 ENSG00000164976.8 KIAA1161 -0.47 2.55E-10 ENSG00000095637.19 SORBS1 -0.48 2.11E-13 ENSG00000163406.9 SLC15A2 -0.48 5.88E-07 ENSG00000279118.1 RP11-517I3.2 -0.48 0.00235446 ENSG00000065361.13 ERBB3 -0.49 6.38E-14 ENSG00000242808.6 SOX2-OT -0.49 1.07E-13 ENSG00000116990.10 MYCL -0.49 0.001905456 ENSG00000169330.7 KIAA1024 -0.50 1.80E-06 ENSG00000253661.1 ZFHX4-AS1 -0.50 6.34E-09 ENSG00000279821.1 RP11- -0.50 0.027757638 1334A24.5 ENSG00000185008.16 ROBO2 -0.50 0.004713852 ENSG00000100427.14 MLC1 -0.50 2.28E-08 ENSG00000197977.3 ELOVL2 -0.50 8.52E-10 ENSG00000183044.10 ABAT -0.50 0.000673888 ENSG00000126878.11 AIF1L -0.50 5.18E-16 ENSG00000164741.13 DLC1 -0.50 1.48E-17 ENSG00000171617.12 ENC1 -0.51 1.50E-28 ENSG00000246763.5 RGMB-AS1 -0.51 1.24E-08 ENSG00000175544.12 CABP4 -0.51 0.003089645 ENSG00000102445.17 KIAA0226L -0.51 8.58E-10 ENSG00000163431.12 LMOD1 -0.51 0.00846078 ENSG00000274813.1 RP11-21I4.3 -0.52 0.030360477 ENSG00000136114.14 THSD1 -0.53 2.94E-05 ENSG00000142149.7 HUNK -0.54 8.68E-06 ENSG00000279802.1 CTA-544A11.1 -0.54 0.019215401

193

ENSG00000134317.16 GRHL1 -0.54 0.001878551 ENSG00000101958.12 GLRA2 -0.54 2.13E-10 ENSG00000144355.13 DLX1 -0.55 1.36E-17 ENSG00000247746.4 USP51 -0.55 0.012507548 ENSG00000078018.18 MAP2 -0.55 1.68E-17 ENSG00000109107.12 ALDOC -0.55 7.08E-08 ENSG00000137801.10 THBS1 -0.55 7.67E-11 ENSG00000139946.8 PELI2 -0.56 3.13E-18 ENSG00000134569.8 LRP4 -0.56 4.66E-12 ENSG00000132855.4 ANGPTL3 -0.56 0.023833751 ENSG00000147145.11 LPAR4 -0.56 2.78E-17 ENSG00000214814.5 FER1L6 -0.57 0.007844128 ENSG00000277151.1 RP11-380B4.3 -0.57 0.004542566 ENSG00000160307.8 S100B -0.57 1.48E-23 ENSG00000124772.10 CPNE5 -0.57 0.000228102 ENSG00000227082.1 CH17-437K3.1 -0.58 0.003089645 ENSG00000159713.9 TPPP3 -0.58 1.12E-08 ENSG00000118523.5 CTGF -0.58 1.61E-15 ENSG00000243244.4 STON1 -0.58 0.0145578 ENSG00000143850.11 PLEKHA6 -0.59 0.012563116 ENSG00000272168.4 CASC15 -0.59 3.25E-12 ENSG00000251615.3 RP11-774O3.3 -0.60 0.034028489 ENSG00000232913.6 PLCE1-AS2 -0.60 0.008767974 ENSG00000220517.2 ASS1P1 -0.60 0.015861307 ENSG00000205336.10 GPR56 -0.61 3.98E-16 ENSG00000278997.1 RP11- -0.62 0.018190802 131M11.2 ENSG00000113578.16 FGF1 -0.62 1.66E-05 ENSG00000146674.13 IGFBP3 -0.62 7.68E-07 ENSG00000104490.16 NCALD -0.62 1.20E-19 ENSG00000164106.6 SCRG1 -0.62 2.44E-05 ENSG00000130055.12 GDPD2 -0.63 7.75E-08 ENSG00000081479.11 LRP2 -0.64 6.99E-07 ENSG00000068078.16 FGFR3 -0.64 6.02E-25 ENSG00000136040.7 PLXNC1 -0.64 9.06E-15 ENSG00000225930.3 DKFZP434L187 -0.65 0.023703729 ENSG00000144485.9 HES6 -0.65 4.63E-24 ENSG00000163017.12 ACTG2 -0.65 7.48E-05 ENSG00000241213.1 RP11-768G7.2 -0.66 0.041505577 ENSG00000182732.15 RGS6 -0.66 0.000991012

194

ENSG00000154330.11 PGM5 -0.66 1.67E-09 ENSG00000231764.7 DLX6-AS1 -0.67 0.00533972 ENSG00000074416.12 MGLL -0.67 0.000123673 ENSG00000072657.7 TRHDE -0.67 2.33E-11 ENSG00000173698.16 GPR64 -0.67 0.004836873 ENSG00000116985.9 BMP8B -0.67 1.91E-07 ENSG00000174804.3 FZD4 -0.67 4.65E-33 ENSG00000237440.7 ZNF737 -0.68 1.56E-10 ENSG00000109063.13 MYH3 -0.68 0.009703388 ENSG00000264630.4 PRKCA-AS1 -0.69 0.023688246 ENSG00000171094.14 ALK -0.69 0.009373226 ENSG00000188039.12 NWD1 -0.70 0.00596925 ENSG00000185046.17 ANKS1B -0.70 5.74E-27 ENSG00000276772.1 CTD- -0.70 0.008732635 2515H24.2 ENSG00000118322.11 ATP10B -0.71 4.16E-10 ENSG00000152689.16 RASGRP3 -0.72 0.039661855 ENSG00000164929.15 BAALC -0.72 2.28E-07 ENSG00000163661.3 PTX3 -0.72 3.79E-28 ENSG00000117834.11 SLC5A9 -0.73 0.011379444 ENSG00000116329.9 OPRD1 -0.73 3.84E-13 ENSG00000170873.17 MTSS1 -0.73 1.90E-20 ENSG00000197928.9 ZNF677 -0.74 7.21E-05 ENSG00000168913.6 ENHO -0.76 0.02416983 ENSG00000140287.9 HDC -0.77 1.10E-07 ENSG00000203506.4 RBMS3-AS2 -0.77 0.000962465 ENSG00000187134.11 AKR1C1 -0.78 1.11E-15 ENSG00000151632.15 AKR1C2 -0.79 3.01E-08 ENSG00000254262.1 RP11-58O3.2 -0.79 0.035856089 ENSG00000144406.17 UNC80 -0.79 1.21E-07 ENSG00000154258.15 ABCA9 -0.80 0.00172682 ENSG00000228842.3 PCDH9-AS2 -0.80 0.00716192 ENSG00000140522.10 RLBP1 -0.81 0.03127903 ENSG00000254672.1 RP5-916O11.3 -0.81 0.032957569 ENSG00000110446.8 SLC15A3 -0.82 0.001569901 ENSG00000280339.1 RP11-736K20.4 -0.83 0.028708774 ENSG00000074181.7 NOTCH3 -0.84 7.52E-07 ENSG00000111783.11 RFX4 -0.84 0.007866219 ENSG00000140022.8 STON2 -0.84 2.93E-72 ENSG00000261150.2 EPPK1 -0.84 0.000720456

195

ENSG00000144668.10 ITGA9 -0.84 3.85E-12 ENSG00000255471.1 RP11-736K20.5 -0.85 0.00759323 ENSG00000203685.8 C1orf95 -0.85 3.10E-07 ENSG00000173258.11 ZNF483 -0.85 1.52E-07 ENSG00000157150.4 TIMP4 -0.87 4.54E-08 ENSG00000153885.13 KCTD15 -0.88 2.25E-07 ENSG00000155629.13 PIK3AP1 -0.89 4.40E-33 ENSG00000158008.8 EXTL1 -0.91 0.000100781 ENSG00000129244.7 ATP1B2 -0.91 0.02073426 ENSG00000168675.17 LDLRAD4 -0.91 3.74E-17 ENSG00000048540.13 LMO3 -0.92 6.92E-12 ENSG00000120162.9 MOB3B -0.93 7.07E-13 ENSG00000138449.9 SLC40A1 -0.93 0.009699982 ENSG00000168772.10 CXXC4 -0.94 1.90E-07 ENSG00000144596.10 GRIP2 -0.99 1.31E-13 ENSG00000280028.1 RP11-118E18.1 -0.99 0.000152758 ENSG00000112902.10 SEMA5A -1.00 3.86E-07 ENSG00000126709.13 IFI6 -1.00 0.001352827 ENSG00000174498.12 IGDCC3 -1.00 2.48E-06 ENSG00000115844.9 DLX2 -1.01 2.20E-05 ENSG00000136160.13 EDNRB -1.01 1.04E-44 ENSG00000144476.5 ACKR3 -1.02 5.67E-26 ENSG00000188171.13 ZNF626 -1.06 4.56E-08 ENSG00000162738.5 VANGL2 -1.06 1.78E-06 ENSG00000171385.8 KCND3 -1.07 2.22E-18 ENSG00000164946.18 FREM1 -1.08 5.67E-08 ENSG00000197921.5 HES5 -1.17 6.81E-05 ENSG00000265843.2 LINC01029 -1.17 0.000190466 ENSG00000007237.17 GAS7 -1.18 1.85E-08 ENSG00000268100.1 ZNF725P -1.20 0.00074153 ENSG00000214652.5 ZNF727 -1.34 4.25E-06 ENSG00000185885.14 IFITM1 -1.36 0.020346496 ENSG00000145569.5 FAM105A -1.36 1.79E-06 ENSG00000078401.6 EDN1 -1.37 1.68E-22 ENSG00000089127.11 OAS1 -1.42 0.012362724 ENSG00000104267.8 CA2 -1.62 3.62E-07 ENSG00000137959.14 IFI44L -1.73 7.82E-05 ENSG00000135114.11 OASL -1.89 0.016523792 ENSG00000165949.11 IFI27 -1.92 3.27E-05 ENSG00000111335.11 OAS2 -2.02 0.003861093

196

ENSG00000183486.11 MX2 -2.30 0.001415233 ENSG00000165029.14 ABCA1 -3.11 8.12E-10

197

CURRICULUM VITAE

Tadas K. Rimkus 4117 Field Crossing Drive Winston Salem, NC 27107 Phone: (443) 896-3052 Email: [email protected]

CAREER OBJECTIVE

Translational research scientist with 8 years of cancer research and project build-out experience. Enthusiastic about clinical drug development, with goal of driving research and clinical strategies in novel cancer therapeutics. Project lead in the development and execution of multiple research projects, with written publication in high-impact journals and meritorious presentations at international cancer research conferences.

EXECUTIVE SUMMARY

As a current graduate research associate at the Wake Forest School of Medicine, I have been an author on eight scientific articles on topics including breast- and neuro-oncology and cancer therapeutics. I have written and developed protocols and SOPs for both pre-clinical and clinical studies in the field of oncology, assuming the role of liaison for animal research regulatory affairs. My experiences are focused in oncology, scientific communication, pre-clinical drug development, and project planning and management. I am an award-winning cancer researcher with established ability in leading pre-clinical research, communicating scientific information, data analysis, and IACUC-mandated monitoring. Strength and experience in:

Doctorate in Cancer Biology Pre-clinical Oncology Studies Regulatory/ICH-GCP Guidelines Clinical Trial Design Project Management Protocols and Amendments Relationship Building Collaborative Research Scientific Strategy

PROFESSIONAL EXPERIENCE

Wake Forest University Health Sciences  Developed, managed, and executed multiple team-oriented research projects focusing on therapy development for solid tumors including breast cancer and glioblastoma, generating funding from external and internal funding sources totalling over $1,000,000.  Cultivated and maintained professional collaborations with key scientists and clinicians, resulting in eight peer-reviewed publications on topics including glioblastoma, breast cancer, and cancer therapeutics.  Developed laboratory SOPs in animal research (orthotopic brain tumor models), molecular biology (patient-derived xenografts, CRISPR/Cas9, immunohistochemistry, cell culture), biochemistry (protein ubiquitination, mass spectrometry) and bioinformatics (clinical data analysis, RNA-seq, R programming).  Assisted in preclinical biomarker research, trial design, and IRB protocol writing that led to approval of a pilot clinical study in breast cancer patients.  Effectively communicated scientific data at national conferences and internal seminars.  Awarded AACR-American Brain Tumor Association Scholar-in-Training Award for meritorious scientific presentation at the 2019 AACR Annual Meeting.  Developed and managed laboratory animal research protocols. Collected, reviewed, and maintained regulatory documents per institutional Animal Research Program (ARP) SOPs.  Mentored and trained six (6) graduate and two (2) post-baccalaureate students  Managed laboratory budget and ordering systems, and implemented laboratory logistics SOPs

Business Development Fellow – Wake Forest Innovations (2017 – 2018)  Analyzed market trends and scientific feasibility related to internal stakeholder technologies

198

 Developed strategic reports of internal stakeholder technologies to facilitate intellectual property development  Attended seminars outlining drug and medical device development processes, from pre- clinical discovery to clinical trials to marketing and regulatory affairs  Streamlined intra-office database for pharmaceutical industry and academia relations to facilitate interactions between Wake Forest University academic labs and industry contacts.  Provided support to case manager on evaluation and development of intellectual property.

Course Instructor – Wake Forest School of Medicine (2018 – present)  Organized and developed curriculum for GRAD715: Career Development in Biomedical Sciences  Mentored 70+ first-year graduate students in career development and networking skills  Moderated speaker panel sessions on Business of Science, Medical Affairs, Science Policy, and Clinical Trial Management  Communicated with industry professionals on a peer-to-peer basis to coordinate panel sessions

Undergraduate Student Researcher – Elon University (2012 – 2014)  Conducted research to study the relationship between cell membrane lipid micro-domains and protein localization and function  Awarded Elon College Fellowship to facilitate independent research  Participated in a journal club focused on protein and cell membrane biophysics  Presented research at internal seminars and external conferences

EDUCATION

Wake Forest University Graduate School of Arts & Sciences, Winston-Salem, NC 05/2019 Ph.D., Cancer Biology

 Characterized novel mechanisms of glioblastoma tumor initiation and progression as part of dissertation studies, which were published twice in Cancer Research.  Awarded AACR-American Brain Tumor Association Scholar-in-Training Award for meritorious abstract at the AACR Annual Meeting 2019

Elon University, Elon, NC 05/2014 Bachelor of Science, Biochemistry (major), Neuroscience (minor); Cum Laude

 Presented research on lipid-anchored protein membrane localization at regional research conference  Earned an Elon College Fellows Independent Research Fellowship to fund undergraduate research project  Earned the J. Mark and Kate Strader McAdams Scholarship in Chemistry for outstanding service to the Chemistry department

HONORS AND AWARDS

AACR-American Brain Tumor Association Scholar-in-Training Award, 2019 Wake Forest University Alumni Travel Award, 2018-2019 Wake Forest University Graduate School of Arts & Sciences Research Fellowship, 2014-2019 Elon University Presidential Scholarship, 2010-2014 Elon College Fellows Independent Research Fellowship, 2010-2014 J. Mark and Kate Strader McAdams Scholarship in Chemistry, 2012-2014 Omicron Delta Kappa, National Leadership Honor Society, 2013-2014 Phi Lambda Upsilon, National Chemistry Honor Society, 2013-2014

PUBLICATIONS

199

1. Rimkus, T., Zhu, D., Carpenter, R.L., Paw, I., Arrigo, A., Routh, E., Sirkisoon, S., Doheny, D., Aguayo, N., Jin, G., Lee, J., Pasche, B., Debinski, W., Lo, HW. NEDD4-mediated degradation of TUSC2 promotes GBM development and progression. In preparation. 2. Sirkisoon, S., Carpenter, R., Rimkus, T., Doheny, D., Zhu, D., Aguayo, N., Xing, F., Chan, M., Ruiz, J., Metheny-Barlow, L., Strowd, R., Masters, A., Lin, J., Thomas, A., Pasche, B., Debinski, W., Watabe, K., Lo, HW. The oncogenic tGLI1 transcription factor is a novel mediator of breast cancer brain metastasis. Submitted to Oncogene. 3. Rimkus, T.*, Carpenter, R.*, Sirkisoon, S., Zhu, D., Pasche, B., Chan, M., Lesser, G., Tatter, S., Watabe, K., Debinski, W., and Lo, HW. Truncated glioma-associated oncogene homolog 1 (tGLI1) mediates mesenchymal glioblastoma via transcriptional activation of CD44. Cancer Research. 78(10): 2589-600 (2018) 4. Sirkisoon, S.*, Carpenter, R.*, Rimkus, T., Anderson, A., Harrison, A., Lange, A., Jin, G., Watabe, K., and Lo, HW. Interaction between STAT3 and GLI1/tGLI1 oncogenic transcription factors promotes the aggressiveness of triple-negative and HER2-enriched breast cancers. Oncogene. 37: 2502-14 (2018). 5. Rimkus, T., Sirkisoon, S., Harrison, A., Lo HW. Tumor suppressor candidate 2 (TUSC2, FUS-1) and human cancers. Discovery Medicine. 23(128): 325-330 (2017). 6. Carpenter, RL., Sirkisoon, S., Zhu, D., Rimkus, T., Harrison, A., Anderson, A., Paw, I., Qasem, S., Xing, F., Liu, Y., Chan, M., Metheny-Barlow, L., Pasche, BC., Debinski, W., Watabe, K., and Lo, HW. Combined inhibition of AKT and HSF1 inhibits breast cancer stem cells and tumor growth. Oncotarget. 8(43): 73947-63 (2017). 7. Rimkus, TK., Carpenter, RL., Qasem, S., Chan, M., and Lo, HW. Targeting the Hedgehog signaling pathway: Review of Smoothened and GLI inhibitors. Cancers. 8(2): 1-23 (2016). 8. Sirkisoon, SR., Carpenter, RL., Rimkus, TK., Miller, L.,Metheny-Barlow, L., and Lo, HW. EGFR and HER2 signaling in breast cancer brain metastasis. Frontiers in Bioscience (Elite Ed). 1(8): p. 245-63 (2016).

PRESENTATIONS

Oral Presentations 1. “Molecular mechanisms of tumor suppressor TUSC2 in glioblastoma,” Section on Brain Tumor Biology and Radiation Oncology. Wake Forest School of Medicine, 2018. 2. “tGLI1 mediates mesenchymal glioblastoma via transcriptional activation of CD44,” Cancer Biology Seminar Series. Wake Forest School of Medicine, 2018. 3. “TUSC2 downregulation and tGLI1 overexpression promote GBM development and progression,” Cancer Biology Seminar Series. Wake Forest School of Medicine, 2017. 4. “tGLI1 is a novel mediator of mesenchymal subtype of glioblastoma and novel transcriptional activator of CD44,” Section on Brain Tumor Biology and Radiation Oncology. Wake Forest School of Medicine, 2017. 5. “TUSC2 downregulation and gain of tGLI1 regulate gliomagenesis and GBM progression,” Section on Brain Tumor Biology and Radiation Oncology. Wake Forest School of Medicine, 2016. 6. “The role of tGLI1 in gliomagenesis,” Cancer Biology Seminar Series. Wake Forest School of Medicine, 2016.

Poster Presentations 1. “Roles of tumor suppressor candidate 2 (TUSC2) in glioblastoma progression and gliomagenesis,” AACR Annual Meeting 2019. Atlanta, GA. 2. “Truncated glioma-associated oncogene homolog 1 (tGLI1) mediates mesenchymal glioblastoma via transcriptional activation of CD44,” AACR Annual Meeting 2018. Chicago, IL. 3. “Truncated glioma-associated oncogene homolog 1 (tGLI1) mediates mesenchymal glioblastoma via transcriptional activation of CD44,” Wake Forest University Graduate Research Day, 2018.

TEACHING RESPONSIBILITIES

200

 Instructor, GRAD715 Career Planning in Biomedical Sciences; Wake Forest School of Medicine, 2018-present  Student Lecturer, Senior Seminar in Biochemistry; Elon University, 2014  Head Teaching Lab Assistant, CHML351, 353 Biochemistry Lab I, II; Elon University, 2013- 2014  Teaching Lab Assistant, CHML211, 212 Organic Chemistry Lab I, II; Elon University, 2012- 2013  Teaching Lab Assistant, CHML111, 112 General Chemistry Lab I, II; Elon University, 2011- 2012

PROFESSIONAL CERTIFICATES & CONTINUING EDUCATION

 CITI ICH-GCP Certification  Making Medicines: The Process of Drug Development (Eli Lilly & Co)  Commercializing Innovation: Principles of Intellectual Property Development

SERVICE AND OUTREACH

Intramural  Executive Committee Member, Peer Mentoring Program, 2017-present  Student Representative, Academic Grievance Committee, 2017-present  MCB Program Representative, Graduate Student Association, 2015-present  Outreach Partner, Brain Awareness Council, 2015-present  Ad Hoc Reviewer – Cancer Research, 2019  Ad Hoc Reviewer – Cancer Letters, 2017

Extramural  Venture Café Ambassador; Winston-Salem, NC, 2017-present  Outreach Volunteer, American Cancer Society; Winston-Salem, NC, 2015-present  Local Volunteer, The Ability Experience; Elon, NC, 2012-2014

Student Mentorship / Training  Austin Arrigo, Wake Forest School of Medicine, Ph.D. in Cancer Biology  Steven Moran, Wake Forest School of Medicine, Ph.D. in Molecular and Cellular Biosciences  Marlyn Anguelov, Wake Forest School of Medicine, MS Thesis in Biomedical Sciences  Sontoria King, Wake Forest School of Medicine, MS in Biomedical Sciences  Daniel Doheny, Wake Forest School of Medicine, MS Thesis in Biomedical Sciences  Alexandria Harrison, Wake Forest School of Medicine, Ph.D. in Molecular Medicine and Translational Science  Eric Altamura, Wake Forest University, Biology Program  Ashley Anderson, Wake Forest School of Medicine, MS Thesis in Biomedical Sciences

MEMBERSHIPS IN PROFESSIONAL SOCIETIES

 American Association for Cancer Research, 2016-present  Radiation Research Society, 2016-present  American Society for Biochemistry and Molecular Biology, 2013-present  American Chemical Society, 2012-present

201